Data Analysis, Data Science, Machine Learning

Data Analysis, Data Science, Machine Learning

General videos

IBM Technology - AI, Machine Learning, Deep Learning and Generative AI Explained

General books

Engineering with Utsav - Books Every Software Engineer Should Read in 2025
PythonProgrammer - I've read 57 Books on AI and Data Science - these are the best (for 2025)

Data Analysis

Data Analysis general (how to learn, advices etc.)

Data Analysis courses

Shashank Kalanithi - Day in the Life of a Data Analyst (playlist)
- Data Analyst - Python PCA Application w/ Streamlit (29K views, 2 years ago, 1:11:54)
- Day in the Life of a Data Analyst - SurveyMonkey Data Transformation (3M views, 2 years ago, 1:17:14)
- Day in the Life of a Data Analyst - Bing Maps API (20K views, 2 years ago, 29:00)
- Day in the Life of a Data Analyst - SurveyMonkey Data Transformation (Using R) (18K views, 2 years ago, 41:55)
- Data Analyst - PyCaret to Enhance Your Machine Learning (4K views, 2 years ago, 31:04)
- Day in the Life of a Data Analyst - Google Data Studio (28K views, 2 years ago, 31:33)
- Making Money (as a data analyst) - Webscraping with ScraperAPI (20K views, 2 years ago, 57:13)
- Data Cleaning CHALLENGE (can you think of a better solution?) (131K views, 2 years ago, 22:41)
Shashank Kalanithi - A Data Crash Course | 100+ Key Data Concepts

Alex The Analyst (playlist) - Data Analyst Bootcamp - course + my notes

Link to a course on YouTube

1. FREE Data Analyst Bootcamp!! (497K views, 6 months ago, 6:52)
2. How to Become a Data Analyst in 2023 (Completely FREE!) (638K views, 7 months ago, 13:59)
3. SQL Basics Tutorial For Beginners | Installing SQL Server Management Studio and Create Tables | 1/4 (893K views, 3 years ago, 9:37)
4. SQL Basics Tutorial For Beginners | Select + From Statements | 2/4 (299K views, 3 years ago, 6:14)

<> is the not equal operator.

My comment: SELECT is a projection operation. It selects the columns that we want to see in the result.
5. SQL Basics Tutorial For Beginners | Where Statement | 3/4 (204K views, 3 years ago, 7:58)
6. SQL Basics Tutorial For Beginners | Group By + Order By Statements | 4/4 (194K views, 3 years ago, 8:09)

COUNT() is not a (regular) column (from a table). It's a derived column/field (and an aggregate function BTW). That's why we don't need to include it in the GROUP BY clause.

ORDER BY can be used not just on one column, but on multiple columns as well. In that case, the order of columns in the ORDER BY clause matters.
7. Intermediate SQL Tutorial | Inner/Outer Joins | Use Cases (315K views, 3 years ago, 15:53)
For inner join, it's not important which ID column we use in the SELECT clause, because they are the same. But for outer join, it's important to use the ID column from the table that we want to keep all the rows from.

Left join keeps all the rows from the left table, and right join keeps all the rows from the right table.

My comment (about the syntax of JOIN; BTW the word "INNER" is optional):
```
SELECT * FROM EmployeeDemographics ed
INNER JOIN EmployeeSalary es
ON ed.EmployeeID = es.EmployeeID
```
is the same as:
```
SELECT * FROM EmployeeDemographics
INNER JOIN EmployeeSalary
ON EmployeeDemographics.EmployeeID = EmployeeSalary.EmployeeID
```
My comment (about for example LEFT (OUTER) JOIN):
```
SELECT *
FROM EmployeeDemographics
LEFT OUTER JOIN EmployeeSalary ON EmployeeDemographics.EmployeeID = EmployeeSalary.EmployeeID
```
This takes everything from the "left" table (EmployeeDemographics) and matches it with the "right" table (EmployeeSalary). If there is no match, it will still take the row from the "left" table, but it will fill the columns from the "right" table with NULLs.

My comment (about emulating FULL OUTER JOIN in MySQL)
```
SELECT *
FROM EmployeeDemographics
LEFT JOIN EmployeeSalary ON EmployeeDemographics.EmployeeID = EmployeeSalary.EmployeeID

UNION

SELECT *
FROM EmployeeDemographics
RIGHT JOIN EmployeeSalary ON EmployeeDemographics.EmployeeID = EmployeeSalary.EmployeeID;
```
8. Intermediate SQL Tutorial | Unions | Union Operator (156K views, 3 years ago, 5:25)
Check out the previous example for FULL OUTER JOIN emulation in MySQL.

JOIN combines both tables based on common column(s) (in our last video that column was the EmployeeID which we had in both tables), while UNION combines both tables based on common rows.

My comments:
- JOIN combines both tables horizontally, while UNION combines both tables vertically. In other words, JOIN puts both tables side by side, while UNION puts both tables on top of each other (below each other).
- Also, UNION removes the duplicates, while UNION ALL keeps the duplicates.
My comment (about UNION and UNION ALL):
```
SELECT *
FROM EmployeeDemographics
UNION
SELECT *
FROM EmployeeSalary;

SELECT *
FROM EmployeeDemographics
UNION ALL
SELECT *
FROM EmployeeSalary;
```
The first query will return 5 rows, while the second query will return 6 rows (because there is one duplicate row in the EmployeeSalary table).
My comment (about the example in the previous video, about FULL JOIN done using UNION):

Even though it looks as though the results is horizontal, it's actually vertical. How? The first query returns 5 rows, while the second query returns 4 rows. So the result of the UNION is 9 rows (4 rows are placed below first 5 rows and then), which is the same as the result of the FULL JOIN.
My comment (about selecting columns in UNION):
```
SELECT EmployeeID, FirstName, Age
FROM EmployeeDemographics
UNION
SELECT EmployeeID, JobTitle, Salary
FROM EmployeeSalary
```
If we select columns like this, we will not get the desired result. Why? Because the columns in the first query are different than the columns in the second query. So we need to select the same number of columns in both queries, and the columns need to be of the same type (or at least compatible types).

9. Intermediate SQL Tutorial | Case Statement | Use Cases (185K views, 3 years ago, 7:26)

Case statement allows us to specify a condition and then also specify what we want returned (a result) if that condition is met. (Case statement allows us to create a new column based on the values of other columns.)

We can put multiple WHEN statements in the CASE statement. Example:
SELECT FirstName, LastName, Age,
CASE
WHEN Age < 18 THEN 'Minor'
WHEN Age >= 18 AND Age <= 65 THEN 'Adult'
ELSE 'Senior'
END AS 'AgeGroup'
FROM EmployeeDemographics;
Something to note is that the CASE statement is evaluated in order. So if the first WHEN statement is true, then it will return that result and not check the other WHEN statements.

My comment:

So if we have multiple WHEN statements, we need to make sure that the most specific WHEN statement is first, and the most general WHEN statement is last. Example:
SELECT FirstName, LastName, Age,
CASE
	WHEN Age > 30 THEN 'Old'
	WHEN Age = 38 THEN 'Stanley' # Previous WHEN statement is true, so this one is not checked. So this one is never true.
	ELSE 'Baby'                  # Solution: change the order of WHEN statements.
END
FROM EmployeeDemographics
WHERE Age is NOT NULL
ORDER BY Age

Another example (about using CASE statement with JOIN):

SELECT *,
CASE
	WHEN JobTitle = 'Salesman' THEN Salary + (Salary * .10)
	WHEN JobTitle = 'Accountant' THEN Salary + (Salary * .05)
	WHEN JobTitle = 'HR' THEN Salary + (Salary * .000001)
END AS 'Salary after raise'
FROM EmployeeDemographics ed
JOIN EmployeeSalary es
ON ed.EmployeeID = es.EmployeeID

10. Intermediate SQL Tutorial | Having Clause (102K views, 3 years ago, 3:31)
The HAVING clause is used to filter the results of an aggregate function. Example:
```
SELECT JobTitle, COUNT(JobTitle)
FROM EmployeeDemographics ed
JOIN EmployeeSalary es
ON ed.EmployeeID = es.EmployeeID
GROUP BY JobTitle
HAVING COUNT(JobTitle) > 1 # This cannot be placed above GROUP BY.
                           # It has to be placed below GROUP BY because it's an aggregate function.
```
The SQL query components need to be in a specific order due to the way SQL engines parse and execute queries. The SQL standard defines a particular sequence in which the clauses should appear in a SELECT statement. Here is the typical order:
Why does HAVING clause have to go below GROUP BY clause?

The SQL query components need to be in a specific order due to the way SQL engines parse and execute queries. The SQL standard defines a particular sequence in which the clauses should appear in a SELECT statement. Here is the typical order:
```
1. SELECT
2. FROM
3. JOIN
4. WHERE
5. GROUP BY
6. HAVING
7. ORDER BY
```
Here's why each step comes where it does:
```
1. SELECT: Specifies the columns you want.
2. FROM: Specifies the tables from which to select or delete or the tables to update.
3. JOIN: Combines rows from two or more tables based on a related column between them.
4. WHERE: Filters records before any groupings are made.
5. GROUP BY: Groups records after the WHERE clause has been applied. The grouping is done on the basis of columns. Aggregate functions (like COUNT, AVG, MAX, etc.) then operate on these groups.
6. HAVING: Filters records after the GROUP BY clause has been applied.
7. ORDER BY: Sorts the records, but does this last, after all filtering and grouping have been done.
```
My comment (a quick recap) of the difference between WHERE and HAVING:
- WHERE filters records before any groupings are made.
- HAVING filters records after the GROUP BY clause has been applied.
My comment about aggregate functions vs. GROUP BY:
- Aggregate functions (e.g., COUNT, SUM, AVG, MIN, MAX): Perform calculations on a set of values and return a single value.
- GROUP BY clause: Groups rows that have the same values in specified columns into summary rows.
11. Intermediate SQL Tutorial | Updating/Deleting Data (86K views, 3 years ago, 4:37)

The difference between INSERT and UPDATE is that INSERT adds a new row to a table, while UPDATE modifies existing rows in a table.
Example:
```
UPDATE EmployeeDemographics
SET Age = 31, Gender = 'Female'
WHERE EmployeeID = 1012
```
Good practice, before DELETE, is to use SELECT to see what we are going to delete.
```
SELECT * FROM EmployeeDemographics
WHERE EmployeeID  = 1005
```
DELETE statement is used to delete rows from a table. Example:
```
DELETE FROM EmployeeDemographics
WHERE EmployeeID  = 1005
```

12. Intermediate SQL Tutorial | Aliasing (85K views, 3 years ago, 6:12)

Useful hint for writing SQL statements:
When writing the SQL statement, first write the FROM clause with an alias. Then, when writing the SELECT statement, use the alias to refer to the table - which will give us a dropdown list of all the columns in that table.```

Now, continuing with the lecture...

Aliasing is used to give a table, or a column in a table, a temporary name.

My example:

SELECT ed.EmployeeID, ed.FirstName, ed.LastName, es.Salary
FROM EmployeeDemographics ed
JOIN EmployeeSalary es
ON ed.EmployeeID = es.EmployeeID

Another example:

SELECT FirstName AS fn
FROM EmployeeDemographics AS ed

is the same as:

SELECT FirstName fn
FROM EmployeeDemographics

Another example:

SELECT CONCAT(FirstName, ' ', LastName) AS 'Full Name'
FROM EmployeeDemographics

Another time we'll use aliasing in the SELECT statement is when we want to use an aggregate function. Example:
SELECT AVG(Age) AS 'Average Age'
FROM EmployeeDemographics

Untill now, we used aliasing for column names. But we can also use aliasing for table names. Example:
SELECT Demographics.EmployeeID
FROM EmployeeDemographics Demographics

13. Intermediate SQL Tutorial | Partition By (163K views, 2 years ago, 4:14)

PARTITION BY statement is often compared to GROUP BY statement. But they are not the same. The GROUP BY statement is going to reduce the number of rows in our result set by actually "rolling them up" (by calculating the sum, or averages for each group), while the PARTITION BY statement is going to divide the result set into partitions (and change how the window function is calculated) without reducing the number of rows in our result set.

SELECT ed.FirstName, ed.LastName, ed.Gender, es.Salary, COUNT(Gender) OVER (PARTITION BY Gender) as TotalGender FROM EmployeeDemographics ed JOIN EmployeeSalary es ON ed.EmployeeID = es.EmployeeID

What is a "window function"? A window function is a function that can be applied to a partition of rows. It is also called an "analytic function". Another way to think about it is that a window function is a function that can be applied to a group of rows that are related to the current row.

ChatGPT's answer:

A window function in SQL allows you to perform calculations across a set of table rows that are related to the current row, almost like a sliding "window" over the data. This is analogous to the way a window in a GUI application displays a subportion of the overall data. What makes window functions unique is that they provide access to more than just the current row of the query result.

Unlike aggregate functions, which return a single result value based on a group of rows, window functions return a single value for each row from the underlying query result based on a window of rows related to that row.

Key components of window functions:

1. OVER() Clause: This defines the window of rows for the function to operate on. You can specify ordering and partitioning within this clause.
2. PARTITION BY: This divides the result set into partitions to which the window function is applied. For instance, if you want to operate on groups of rows having the same value in a certain column (e.g., date or category), you'd use PARTITION BY.
3. ORDER BY: Within the OVER() clause, this determines the order in which rows will be processed by the window function.
4. ROWS/RANGE: Specifies which rows are included in the frame. The frame is a subset of the current partition and is used for calculation in functions like running totals.

Examples of window functions:

- Ranking functions: ROW_NUMBER(), RANK(), DENSE_RANK(), NTILE().
- Analytic functions: LEAD(), LAG(), FIRST_VALUE(), LAST_VALUE().
- Aggregate functions: SUM(), AVG(), MIN(), MAX(). These can be used as window functions when combined with the OVER() clause.

Using our earlier example, COUNT(Gender) OVER (PARTITION BY Gender) is a window function. For each row in the dataset, it counts the number of rows with the same Gender value.

In essence, window functions let you perform calculations that require considering a range or "window" of rows relative to the current row, without collapsing all those rows into a single output row.

Aditional explanation Colt Steele - SQL Window Functions in 10 Minutes

Several things to note:

Window functions perform aggregate operations on group of rows but they produce a result FOR EACH ROW. So we have individual row data alongside aggregated data.

The OVER() clause constructs a window. When it's empty, the window will include all records.

Use ORDER BY inside of the OVER() clause to re-order rows within each window.

The PARTITION BY clause divides the window into smaller sets or partitions. The window function is applied to each partition separately and computation restarts for each partition. (Copilot suggestion)

14. Advanced SQL Tutorial | CTE (Common Table Expression) (214K views, 2 years ago, 3:44)

CTE is a common table expression and it's a named temporary result set which is used to manipulate (Copilot suggestion: the data in the result set further) the complex sub-queries data. This only exists within the scope of the statement that we're about to write. Once we cancel out of this query it's like it never existed. A CTE is also only created in memory rather than a tempdb file like a temp table would be, but in general a CTE acts very much like a subquery and so if we know how to write subqueries we can easily learn how to write CTEs.
CTEs are sometimes called "WITH" queries because we use the WITH keyword to create them. The syntax is as follows:
```
WITH CTE_Employee as
(SELECT ed.FirstName, ed.LastName, ed.Gender, es.Salary,
COUNT(Gender) OVER (PARTITION BY Gender) as TotalGender,
AVG(Salary) OVER (PARTITION BY Gender) AS AvgSalary
FROM EmployeeDemographics ed
JOIN EmployeeSalary es
ON ed.EmployeeID = es.EmployeeID
WHERE Salary > '45000')

SELECT *
FROM CTE_Employee
```
Explanation suggested by the Copilot: CTE is a temporary result set that we can reference within another SQL statement. It's similar to a subquery, but it's more readable and easier to maintain.

Example:
```
WITH EmployeeCTE AS (
	SELECT *
	FROM EmployeeDemographics
	WHERE Age > 30
)

SELECT *
FROM EmployeeCTE
```
Aditional explanation (check out LearnatKnowstar - SQL | Subquery or CTE - Which one to choose? Difference between Subquery and CTE):

Excerpt:
CTEs are named queries (suggested by Copilot: temporary result sets) that you can reference (multiple times if needed) within another SQL statement. They are similar to subqueries, but they are more readable and easier to maintain. They are also more flexible than subqueries. For example, you can reference a CTE multiple times in the same SELECT statement, while you can only reference a subquery once. In terms of performance and in terms of how the logic is executed, they are similar to subqueries (no difference as far as I can see at the moment).
There are scenarios where we we can use only a CTE or only a subquery:
- CTE can be recursive, while subquery cannot be recursive:
  - Recursive CTEs can reference themselves in a query. (Suggested by Copilot: This is useful for querying hierarchical data (like a tree structure) or data that has a parent-child relationship.)
- Corelated subqueries can reference outer table from the subquery, while CTE cannot reference outer table from the CTE.
Suggested by Copilot: there are also scenarios where we can use both. In those cases, we should use a CTE because it's more readable and easier to maintain.
15. Advanced SQL Tutorial | Temp Tables (158K views, 2 years ago, 10:19)

Suggested by Copilot: Temp tables are like regular tables, but they are only available to the current session. They are also stored in tempdb, which is a system database. Temp tables are automatically deleted when the session that created them is closed.
We can hit off of this temp table multiple times which we cannot do with something like CTE or subquery, where we can only use it one time or with the subquery where we need to write it multiple times within the query. (TODO: check this)
Example:
```
CREATE TEMPORARY TABLE temp_Employee (
EmployeeID int,
JobTitle varchar(100),
Salary int
)

SELECT *
FROM temp_Employee
```
Maybe the best way to work with temp tables is to take a subset of data from a (much) larger table, put it into a temp table, and then work with that temp table.
Example:
```
INSERT INTO temp_Employee
SELECT *
FROM EmployeeSalary AS es
WHERE es.Salary > '40000'
```
Realistic example: We have a table with 100 million rows. We want to do some analysis on that table, but we don't want to do it on the entire table. So we can create a temp table with a subset of data from that table, and then do the analysis on that temp table. (Suggested by Copilot)
```
CREATE TEMPORARY TABLE temp_Employee2
SELECT JobTitle, Count(JobTitle), Avg(Age), Avg(Salary)
FROM EmployeeDemographics AS ed
JOIN EmployeeSalary AS es
ON ed.EmployeeID = es.EmployeeID
GROUP BY JobTitle
```
Useful for later: A lot of times these temp tables are used in stored procedures. (we'll need to remove the temp table before we can create it again)
Example:
```
DROP TABLE IF EXISTS temp_Employee
CREATE TEMPORARY TABLE temp_Employee (
...
)

INSERT INTO temp_Employee
SELECT *
FROM ...

SELECT *
FROM temp_Employee
```
16. Advanced SQL Tutorial | String Functions + Use Cases (105K views, 2 years ago, 13:49)
17. Advanced SQL Tutorial | Stored Procedures + Use Cases (244K views, 2 years ago, 6:15)
18. Advanced SQL Tutorial | Subqueries (233K views, 2 years ago, 8:37)
19. Data Analyst Portfolio Project | SQL Data Exploration | Project 1/4 (1.3M views, 2 years ago, 1:17:09)
20. Data Analyst Portfolio Project | Data Cleaning in SQL | Project 3/4 (281K views, 2 years ago, 54:44)
21. Pivot Tables in Excel | Excel Tutorials for Beginners (333K views, 1 year ago, 17:35)
22. Formulas in Excel | Excel Tutorials for Beginners (160K views, 1 year ago, 33:54)
23. XLOOKUP in Excel | Excel Tutorials for Beginners (98K views, 1 year ago, 18:47)
24. Conditional Formatting in Excel | Excel Tutorials for Beginners (81K views, 1 year ago, 20:59)
25. Charts in Excel | Excel Tutorials for Beginners (60K views, 1 year ago, 15:11)
26. Cleaning Data in Excel | Excel Tutorials for Beginners (280K views, 1 year ago, 21:04)
27. Full Project in Excel | Excel Tutorials for Beginners (417K views, 1 year ago, 40:50)
28. How to Install Tableau and Create First Visualization | Tableau Tutorials for Beginners (402K views, 1 year ago, 17:04)
29. How to use Calculated Fields and Bins in Tableau | Tableau Tutorials for Beginners (118K views, 1 year ago, 6:25)
30. How to Create Visualizations in Tableau | Tableau Tutorials for Beginners (95K views, 1 year ago, 14:05)
31. How to use Joins in Tableau | Tableau Tutorials for Beginners (69K views, 1 year ago, 14:29)
32. Full Beginner Project in Tableau | Tableau Tutorials for Beginners (154K views, 1 year ago, 44:18)
33. How to Install Power BI | Building First Visualization | Microsoft Power BI for Beginners (155K views, 1 year ago, 12:50)
34. How to use Power Query in Power BI | Microsoft Power BI for Beginners (107K views, 1 year ago, 13:07)
35. How to Create and Manage Relationships in Power BI | Microsoft Power BI for Beginners (74K views, 1 year ago, 8:36)
36. How to use DAX in Power BI | Microsoft Power BI for Beginners (62K views, 1 year ago, 15:44)
37. How to use Drill Down in Power BI | Microsoft Power BI for Beginners (61K views, 1 year ago, 6:02)
38. How to use Conditional Formatting in Power BI | Microsoft Power BI for Beginners (49K views, 1 year ago, 9:53)
39. How to use Bins and Lists in Power BI | Microsoft Power BI for Beginners (32K views, 11 months ago, 9:31)
40. Popular Visualizations in Power BI | Microsoft Power BI for Beginners (35K views, 11 months ago, 14:14)
41. Full Power BI Guided Project | Microsoft Power BI for Beginners (186K views, 11 months ago, 42:37)
42. Installing Jupyter Notebooks/Anaconda | Python for Beginners (105K views, 10 months ago, 10:03)
43. Variables in Python | Python for Beginners (36K views, 10 months ago, 13:17)
44. Data Types in Python | Python for Beginners (33K views, 9 months ago, 21:58)
45. Comparison, Logical, and Membership Operators in Python | Python for Beginners (17K views, 9 months ago, 7:15)
46. If Else Statements in Python | Python for Beginners (16K views, 9 months ago, 6:40)
47. For Loops in Python | Python for Beginners (18K views, 9 months ago, 9:17)
48. While Loops in Python | Python for Beginners (15K views, 9 months ago, 5:40)
49. Functions in Python | Python for Beginners (18K views, 8 months ago, 12:44)
50. Converting Data Types in Python | Python for Beginners (15K views, 8 months ago, 6:36)
51. Building a BMI Calculator with Python | Python Projects for Beginners (29K views, 7 months ago, 14:23)
52. Building an Automated File Sorter in File Explorer using Python | Python Projects for Beginners (22K views, 6 months ago, 16:51)
53. Inspecting Web Pages with HTML | Web Scraping in Python (13K views, 2 months ago, 5:55)
54. BeautifulSoup + Requests | Web Scraping in Python (14K views, 2 months ago, 6:58)
55. Find and Find_All | Web Scraping in Python (9K views, 1 month ago, 12:10)
56. Scraping Data from a Real Website | Web Scraping in Python (42K views, 1 month ago, 25:23)
57. Reading in Files in Pandas | Python Pandas Tutorials (31K views, 6 months ago, 19:17)
58. Filtering Columns and Rows in Pandas | Python Pandas Tutorials (20K views, 5 months ago, 11:49)
59. Indexes in Pandas | Python Pandas Tutorials (13K views, 5 months ago, 11:22)
60. Group By and Aggregate Functions in Pandas | Python Pandas Tutorials (14K views, 4 months ago, 11:05)
61. Merging DataFrames in Pandas | Python Pandas Tutorials (18K views, 3 months ago, 22:09)
62. Creating Visualizations using Pandas Library | Python Pandas Tutorials (18K views, 3 months ago, 16:50)
63. Data Cleaning in Pandas | Python Pandas Tutorials (49K views, 3 months ago, 38:37)
64. Exploratory Data Analysis in Pandas | Python Pandas Tutorials (28K views, 2 months ago, 32:13)
65. Amazon Web Scraping Using Python | Data Analyst Portfolio Project (177K views, 2 years ago, 47:14)
66. Automating Crypto Website API Pull Using Python | Data Analyst Project (35K views, 1 year ago, 51:14)
67. How to Create a Portfolio Website for FREE (322K views, 2 years ago, 35:29)
68. Create the Perfect Data Analyst Resume | Free Templates! (78K views, 5 months ago, 17:37)
69. Top 3 Tips on Using LinkedIn to Land a Job (198K views, 2 years ago, 6:50)
70. How To Download Your Data Analyst Bootcamp Certification (Congrats!!) (25K views, 7 months ago, 1:41)

Data Analysis projects

Data Analysis books

Python for Data Analysis: Data Wrangling with pandas, NumPy, and Jupyter - Wes McKinney (2022, 3rd edition, 579 pages, O'Reilly Media, 5.4 stars from 241 ratings on Amazon)

Data Analysts - what do they do?

Agatha - What I actually do as a Data Analyst | salary, job, harsh reality

Data Science

Data Science general (roadmaps, how to learn, advices etc.)

Gregg Hogg - Data Science Has Changed - Here's What to Do
Sundas Khalid - How to Become a Data Scientist in 2024? (complete roadmap)
Tani Tech - HOW To BECOME A DATA SCIENTIST In 2024
Power Couple - FASTEST Way to Learn Data Science and ACTUALLY Get a Job
Thu Vu data analytics - 👩🏻‍💻 How to learn Data Science FASTER
DataNash - How I'd Learn Data Science In 2023 (If I Could Restart) | A Beginner's Roadmap
DataNash - How I'd Learn Data Science In 2024 (If I Could Restart) - The Ultimate Roadmap
Data Appreciation Society - How To Get a Job in MACHINE LEARNING without a DEGREE | 7 Best Tips (the following URL's are mostly generated using the Copilot, but comments are added by me based on the info from this video)
- Learn Data Exploration (data preprocessing, data cleaning, data visualization)
  - Article on data pre-processing/data exploration
  - Another article explaining why it is so important to know for data scientists and machine learning engineers
- Learn Pandas, SciKit-Learn, NumPy (also Matplotlib, Seaborn)
- SQL ("you only need basic SQL" hmmm)
- Start with startups
- Look for a mentor
  - Data Science Mentors
  - Follow on LinkedIn: Kate Strachnyi, Cassie Kozyrkov, Andrew Ng, Jason Brownlee, Elena Tej Grewal, Vin Vashishta, Kevin Gray
- Personal projects
  - DataCamp - 5 Data Science Projects That Will Get You Hired in 2021

Data Science courses

Data Science projects

Ken Jee - 5 Impactful Data Science Projects For Your Portfolio
- NASA
  - Datasets for NASA missions, research, activities, and more
  - Life science data (from NASA)

Data Science mathematics

Data Science books

Essential Math for Data Science: Take Control of Your Data with Fundamental Linear Algebra, Probability, and Statistics - Thomas Nield (2022, 1st edition, 347 pages, O'Reilly Media, 4.5 stars from 231 ratings on Amazon)
Practical Statistics for Data Scientists: 50+ Essential Concepts Using R and Python - Andrew Bruce, Peter Bruce (2020, 2nd edition, 360 pages, O'Reilly Media, 4.6 stars from 793 ratings on Amazon)
Becoming a Data Head: How to Think, Speak, and Understand Data Science, Statistics, and Machine Learning - Alex J. Gutman, Jordan Goldmeier (2021, 1st edition, 272 pages, Wiley, 4.6 stars from 311 ratings on Amazon)
The Art of Statistics: How to Learn from Data - David Spiegelhalter (2021, 1st edition, 448 pages, Basic Books, 4.6 stars from 3,508 ratings on Amazon)
Data Science from Scratch: First Principles with Python - Joel Grus (2019, 2nd edition, 403 pages, O'Reilly Media, 4.4 stars from 700 ratings on Amazon)
Think Like a Data Scientist: Tackle the data science process step-by-step - Brian Godsey (2017, 1st edition, 328 pages, Manning, 4.5 stars from 40 ratings on Amazon)

Data Scientists - what do they do?

Tani Tech - https://www.youtube.com/watch?v=iymwofogLCs

Machine Learning

Machine learning general (how to learn, advices etc.)

Machine learning projects

Smitha Kolan - Machine Learning Engineer - ABSOLUTE BEGINNER Machine Learning NLP Project (Tutorial)
Boris Meinardus - How To Build A Machine Learning Portfolio in 2024
Boris Meinardus - This ML Project Gives You an Unfair Advantage
- Reimplementing a paper and recreating it's results
  - Step 1: Read the paper
    - Methodology: Use colored markers to highlight the important parts
      - Color no. 1 (General understanding): We want to highlight the parts that are important for understanding the paper
      - Color no. 2 (Implementation details): Actively highlight the parts relevant for the implementation.
        This includes things like used loss loss functions, used activation functions, used architectures, used datasets, used hyperparameters etc. When highlighting these parts, we can already start envisioning how we would start to implement the architecture and which other repositories we alredy know that have similar implementations that we can use for our own new implementation. Or we write down a list of common modules we need to look up existing implementations for. The idea is not to not to reinvent the wheel, but to use existing implementations and to understand how they work and how they can be adapted to our own needs.
      - Color no. 3 (Datasets): Since we are not only reimplement the paper but also recreate results to actually verify whether we implemented the paper successfully, we need to train on the same datasets and evaluate on the same ones. Highlighting those should not too difficult, since the paper should already mention the datasets used and the evaluation metrics used.
        It might also be important to look for details about the compute that the authors used. If they used expensive compute, we might need to look for alternatives or we might need to look for ways to reduce the compute needed. This might be important for us, since we might not have access to the same compute as the authors of the paper.
        Tackle the datasets. That means - find the datasets used in the paper and download them.
  - Step 2: Implementing the dataset and data loader classes (PyTorch).
    Data loader is used to provide data in a preprocessed format. That way we can very consistently get a batch of data and just focus on the training logic. When implementing the dataset and data loader classes he recomments to use Jupyter Notebooks. That makes it very easy to save certain variables (like the path to the dataset) and visualize everything we need.
    One final thing is if we are working on a reinforcement learning paper, the equivalent to implementing the dataset classes is setting up the gym environment.
  - Step 3: Start coding
    (NOTE: Perhaps our paper focuses more on a new training technique rather a new architecture).
    Since we've highlighted the novel elements of the architecture and found existing implementations of modules we might need, his recommendation is to just start coding(!!!). Don't think too much about making it right the first time. Start writing our torch.nn.module class and initialize all layers we might need. If they are larger, more complex models, we can directly start writing a function for those. [TODO: Check the mentioned example].
    - Use "Weights and Biases" to track our experiments. It's a tool that allows us to track our experiments and visualize them. It's a very powerful tool and it's free to use. (added by the Copilot)
- Algorithmic Simplicity - Deriving the Transformer Neural Network from Scratch #SoME3

Diffusion models

DeepFindr - Diffusion models from scratch in PyTorch

Machine Learning basics

David MacKay - Information Theory, Inference, and Learning Algorithms

Produced by: David MacKay (University of Cambridge) Author: David MacKay, University of Cambridge

A series of sixteen lectures covering the core of the book "Information Theory, Inference, and Learning Algorithms" (Cambridge University Press, 2003, http://www.inference.eng.cam.ac.uk/mackay/itila/) which can be bought at Amazon (http://www.amazon.co.uk/exec/obidos/ASIN/0521642981/davidmackay0f-21), and is available free online (http://www.inference.eng.cam.ac.uk/mackay/itila/).

A subset of these lectures used to constitute a Part III Physics course at the University of Cambridge. The high-resolution videos and all other course material can be downloaded from the Cambridge course website (http://www.inference.eng.cam.ac.uk/mackay/itprnn/).

Snapshots of the lecture can be found here: http://www.inference.eng.cam.ac.uk/itprnn_lectures/

These lectures are also available at http://videolectures.net/course_information_theory_pattern_recognition/ (synchronized with snapshots and slides)
1. Lecture 1: Introduction to Information Theory (329K views, 9 years ago, 1:01:51)
2. Lecture 2: Entropy and Data Compression (I): Introduction to Compression, Inf.Theory and Entropy (90K views, 9 years ago,51:09)
3. Lecture 3: Entropy and Data Compression (II): Shannon's Source Coding Theorem, The Bent Coin Lottery (55K views, 9 years ago, 51:01)
4. Lecture 4: Entropy and Data Compression (III): Shannon's Source Coding Theorem, Symbol Codes (40K views, 9 years ago, 56:58)
5. Lecture 5: Entropy and Data Compression (IV): Shannon's Source Coding Theorem, Symbol Codes (33K views, 9 years ago, 1:02:48)
6. Lecture 6: Noisy Channel Coding (I): Inference and Information Measures for Noisy Channels (27K views, 9 years ago, 54:42)
7. Lecture 7: Noisy Channel Coding (II): The Capacity of a Noisy Channel (26K views, 9 years ago, 46:54)
8. Lecture 8: Noisy Channel Coding (III): The Noisy-Channel Coding Theorem (23K views, 9 years ago, 1:08:59)
9. Lecture 9: A Noisy Channel Coding Gem, And An Introduction To Bayesian Inference (I) (15K views, 9 years ago, 48:36)
10. Lecture 10: An Introduction To Bayesian Inference (II): Inference Of Parameters And Models (19K views, 9 years ago, 1:15:53)
11. Lecture 11: Approximating Probability Distributions (I): Clustering As An Example Inference Problem (15K views, 9 years ago, 56:56)
12. Lecture 12: Approximating Probability Distributions (II): Monte Carlo Methods (I) (22K views, 9 years ago, 1:23:48)
13. Lecture 13: Approximating Probability Distributions (III): Monte Carlo Methods (II): Slice Sampling (14K views, 9 years ago, 1:47:57)
14. Lecture 14: Approximating Probability Distributions (IV): Variational Methods (18K views, 9 years ago, 46:35)
15. Lecture 15: Data Modelling With Neural Networks (I): Feedforward Networks: The Capacity Of A Neuron (14K views, 9 years ago, 1:27:16)
16. Lecture 16: Data Modelling With Neural Networks (II): Content-Addressable Memories And State (10K views, 9 years ago, 1:36:33)
r2d3 - A visual introduction to machine learning (text)
- Part 1 - A Decision Tree (text)
- Part 2 - Model Tuning and the Bias-Variance Tradeoff (text)

Machine learning courses

Premium (Udemy, Coursera etc.)

Free

Machine learning books

Machine Learning with Python Cookbook: Practical Solutions from Preprocessing to Deep Learning - Chris Albon (2023, 2nd edition, 413 pages, O'Reilly Media)
Machine Learning with PyTorch and Scikit-Learn: Develop machine learning and deep learning models with Python - Sebastian Raschka, Yuxi (Hayden) Liu (2022, 1st edition, 774 pages, Packt Publishing)
Introduction to Machine Learning with Python: A Guide for Data Scientists - Andreas C. Müller, Sarah Guido (2016, 1st edition, 398 pages, O'Reilly Media)

Deep Learning

Tutorials

Artificial intelligence

Artificial intelligence general (how to learn, advices etc.)

Tutorials

Fine-tuning

Fine-tuning projects

Mathematics for AI

Artificial intelligence articles

WaitButWhy - The AI Revolution: The Road to Superintelligence

Artificial intelligence books

Deep Learning for Coders with Fastai and PyTorch: AI Applications Without a PhD - Jeremy Howard, Sylvain Gugger (2020, 1st edition, 621 pages, O'Reilly Media, 4.7 stars from 500 reviews on Amazon)
Inside Deep Learning: Math, Algorithms, Models - Edward Raff (2022, Annotated edition, 600 pages, Manning, 4.7 stars from 24 reviews on Amazon)
Deep Learning with Python, Second Edition - François Chollet (2021, 2nd edition, 504 pages, Manning, 4.7 stars from 345 reviews on Amazon)
Deep Learning with PyTorch: Build, train, and tune neural networks using Python tools - Eli Stevens, Luca Antiga, Thomas Viehmann (2020, 1st edition, 520 pages, Manning, 4.4 stars from 119 reviews on Amazon)

Mathematics courses

Khan Academy - Statistics and Probability
- Lesson 1: Analyzing one categorical variable
  - Median - is the middle value in a set of data. Example, 100 is median in this array of data: 75, 80, 100 (this is median), 100, 100.
  - Midrange - is the mean of the highest and lowest values in a set of data.
    - Example: for the array of data 75, 80, 100, 100, 100 the midrange is (75+100)/2=87.5
  - Mean - is the average value in a set of data. Example, 100 is mean in this array of data
  - Average - is the average value in a set of data.
    - NOTE: If someone says "average" without giving more information they're probably thinking about the arithmetic mean.
    - Example: for the array of data 75, 80, 100, 100, 100 the average is (75+80+100+100+100)/5=91
  - Mode - is the most commommon score. In 75, 80, 100, 100, 100, the mode is 100 because it appears 3 times.
  - Range - is the difference between the highest and lowest values in a set of data. Example: for the array of data 75, 80, 100, 100, 100 the range is 100-75=25
- Lesson 2
  - Dot plots - are used to show individual data points (bad if we have a lot of data).
  - Bar graphs - are used to show groups of data (good if we have a lot of data).
  - Histograms - displays ranges of data (good if we have a lot of data). Example: 0-10, 11-20, 21-30, etc.

Databases

Interactive SQL tutorials

Platforms for learning

Elements of AI - free course on AI

Talks

Interviews

YouTube channels

metacodeM - Machine Learning, Data Science, Python, and more

Whether you're exploring the realms of AI, data science, machine learning, deep learning, LLM, or statistics, our expert instructors, who are either actively working in tech fields or hold Ph.D.s in their domains, are here to guide you for free.

Our goal is to empower you with the knowledge and skills needed to thrive in the dynamic world of rapidly evolving technologies.

Additionally, we're excited to announce the upcoming launch of our community website, where you can engage with fellow learners, stay updated on industry trends, and access exclusive resources related to Industry 4.0.

We’ll also offer comprehensive courses designed to provide you with in-depth learning experiences. Upon completion, you'll have the opportunity to earn certificates, validating your expertise.

deeplizard - ML + DL + AI tutorials, AI art tutorials, NLP tutorials etc.
Machine Learning Street Talk
DeepFindr - Machine Learning / Data Science topics
Make Data Useful - Data science and analytics channel with a nice mix of Python tutorials and analysis.

Name		Name	Last commit message	Last commit date
Latest commit History 57 Commits
README.md		README.md

mladen/practicing-da-ds-ml-ai

Folders and files

Latest commit

History

Repository files navigation

Data Analysis, Data Science, Machine Learning

General videos

General books

Data Analysis

Data Analysis general (how to learn, advices etc.)

Data Analysis courses

Alex The Analyst (playlist) - Data Analyst Bootcamp - course + my notes

Data Analysis projects

Data Analysis books

Data Analysts - what do they do?

Data Science

Data Science general (roadmaps, how to learn, advices etc.)

Data Science courses

Data Science projects

Data Science mathematics

Data Science books

Data Scientists - what do they do?

Machine Learning

Machine learning general (how to learn, advices etc.)

Machine learning projects

Diffusion models

Machine Learning basics

Machine learning courses

Premium (Udemy, Coursera etc.)

Free

Machine learning books

Deep Learning

Tutorials

Artificial intelligence

Artificial intelligence general (how to learn, advices etc.)

Tutorials

Fine-tuning

Fine-tuning projects

Mathematics for AI

Artificial intelligence articles

Artificial intelligence books

Mathematics courses

Databases

Interactive SQL tutorials

Platforms for learning

Talks

Interviews

YouTube channels

Other links and playlists

Interesting videos involving AI

About

Resources

Stars

Watchers

Forks

Releases

Packages 0

Packages