TA’s Session TDS – 2024 09 29 19 49 IST – Recording#

Duration: 1h 40m

Here’s an FAQ based on the provided TDS live tutorial:

Q1: How do I know if my peer review is an outlier?

A1: You won’t know this directly. When you submit your review, the system automatically checks it against rubrics and other reviews. If your score is significantly different from the others for the same submission (e.g., you give a 0 when others give 80-90, or vice versa), it will be flagged as an outlier.

Q2: Will the bonus marks for peer reviews be calculated automatically?

A2: Yes, the system calculates these bonus marks at the backend. If none of your reviews are flagged as outliers, you’ll receive one bonus mark.

Q3: An email mentioned a third project involving an LLM app. Is that correct?

A3: There are only two main projects for this term (Project 1 and Project 2). The LLM app is actually an opportunity to earn bonus marks, not a full project. You can find more details about this on the Discourse forum.

Q4: How valuable are these bonus marks, and how do they relate to graded assignments (GAs)?

A4: Bonus marks are very valuable. To give you an idea, earning 5 bonus marks is equivalent to getting 5 graded assignments right. This can significantly impact your final grade, especially for difficult parts of the course like ROE. Don’t underestimate their worth!

Q5: Will there be mock ROEs, or will the pattern be similar to past terms?

A5: There isn’t a traditional “mock” ROE. You should look at past ROEs as examples. However, the pattern of ROE questions is never the same from term to term; they are often wildly different. The key is to master the core skills, not a specific pattern.

Q6: What are the core skills I need to master for ROE?

A6: For ROE, you need to be very comfortable with:

• Scraping HTML and PDFs.
• Extracting data from various sources, including databases.
• Using libraries like Pandas and Nominatim.
• Understanding and using CSS selectors (this is crucial!).
• Dealing with authentication when accessing data.
• These skills are often covered in the “optional” videos in Week 2, which you should not ignore, as they are fundamental for ROE.

Q7: How many modules are there in this course?

A7: You can find the full week-by-week breakdown on the course introduction page. It details exactly what will be covered each week.

Q8: Can you give a quick demo of how the LLM app bonus works?

A8: (Instructor demonstrates prompting ChatGPT to create a Python calculator app.) The idea is to use an LLM (like ChatGPT) to generate an application. For the bonus, you’ll:

1. Prompt: Ask the LLM to create an app (e.g., “Create a calculator app in Python”).
2. Copy Code: Take the code generated by the LLM.
3. Run & Debug: Paste it into a development environment (like VS Code), run it, and if there are errors, feed the error message back into the LLM to get it to fix the code. You are not allowed to manually edit the code yourself.
4. Deploy: Get the LLM to provide instructions for deploying the app (e.g., using Streamlit, GitHub Pages, Heroku, Netlify, or Vercel).
5. Record: Record your screen demonstrating the entire process (prompting, running, debugging with LLM, deploying) in a single, unedited 10-minute video. This shows how LLMs can automate tasks like creating apps or cleaning data.

Q9: Where do I upload or showcase my LLM bonus project?

A9: You need to post it in the specific Discourse thread dedicated to the LLM bonus. You can host your app on platforms like GitHub Pages, Heroku, Netlify, or Vercel. If you can’t host it live, you can upload it as a zip file, but this might result in slightly lower marks. Read the pinned Discourse post carefully for all instructions.

Q10: Are there any specific video requirements for the LLM bonus submission?

A10: Yes, your screen recording must be a single, unedited 10-minute video. No editing, trimming, enhancing, or annotating the video is allowed. You should record yourself creating the entire app from start to finish, using only the LLM.

Q11: What is the deadline for the LLM bonus project?

A11: The deadline for the LLM bonus is October 7th. You’ll find this date specified in the Discourse post.

Q12: Is a simple web calculator considered an acceptable app for the LLM bonus?

A12: The criterion is “how impressive the app is.” While a web calculator might be simple, the goal is to see your ingenuity. The project is open-ended, so you decide what kind of app you want to create using the LLM. It should be interactive and accessible via a browser.

Q13: I have no prior programming experience. How can I learn everything quickly enough?

A13: The key is “time management” and “being honest with yourself.” While it might seem like a lot, the course provides all the learning resources. It’s about being proactive. Read all the provided materials, including the “optional” sections and hyperlinks, as they contain valuable information. If you’re new, you’ll need to dedicate more time to practice.

Q14: What kind of data and tools will be used in the projects?

A14: Projects typically involve large, unclean datasets. You’ll be given a set of questions (around 10-15) to answer by processing this data. This often involves using a mix of the tools covered in the course (e.g., HTML/PDF scraping, data extraction from databases, data cleaning, visualization). Sometimes, the data might come in various formats (e.g., Parquet files, SQL databases, PDFs), requiring you to combine information from multiple sources.

Q15: What kind of questions can I expect in ROE?

A15: ROE questions are highly varied and unpredictable; they are “completely different beasts” each term. You cannot simply eyeball an answer. They test your core skills and intuition. For instance, past ROEs have involved:

• Scraping ball-by-ball cricket commentary (including handling programmatic elements that required scrolling to reveal data).
• Extracting and cleaning data from deliberately unclean PDFs.
• Combining information from three different data sources (Parquet, SQL, PDF) to answer complex questions. The key is to master the core skills and be prepared for anything.

Q16: Is ROE open book?

A16: Yes, ROE is completely open book and not proctored. You can use any resources, including ChatGPT, friends, or the internet. However, this highlights that the exam tests your intuition and ability to quickly apply your knowledge, not just memorization. You won’t have time to search for every answer; you’ll need to know the core skills well enough to execute efficiently.

Q17: I don’t have Excel. Do I need to purchase it for the course?

A17: No, you don’t need to purchase Excel. You can use the free cloud version of Microsoft 365, which is sufficient for most tasks. The course also introduces powerful 3D visualization tools within Excel that can create impressive visuals.

Q18: Will JavaScript/CSS skills be required for the projects or ROE?

A18: Yes, especially for ROE. While you’re not expected to be an expert developer, you must be very comfortable with CSS selectors. Week 2 covers this. Understanding CSS selectors is critical for web scraping and extracting specific data elements from web pages.

Q19: I’m facing an issue with GA0, where the modified date of the files changes, causing errors. How can I fix this?

A19: This is a known Windows-specific issue that occurs when using the built-in “extract all” feature for zip files. To resolve this, use the unzip command in Git Bash instead of Windows’ default extractor. This should preserve the original timestamps.

Q20: What is the purpose of refreshing GA0 questions every time I open it?

A20: The instructor’s objective for GA0 (and GA1) is for you to complete all questions in one sitting. This challenges you to plan and execute efficiently. While there have been suggestions for features like auto-save or allowing longer completion times, the current setup is designed to test your ability to tackle the assignment holistically.

Q21: What kind of questions will be in Project 1?

A21: Project 1 typically involves a large, potentially unclean dataset. You’ll be given a set of 10-15 questions to answer by processing and analyzing this data. You’ll need to create charts and present your findings. Depending on the dataset, you might need to use skills like HTML/CSS/JavaScript for data extraction.

Q22: Will the projects require HTML, CSS, or JavaScript?

A22: It depends on the specific dataset provided for the project. Last term, the dataset required web scraping from an HTML source, including dealing with deliberately “noisy” HTML. So, yes, you should be prepared to use your HTML, CSS (for selectors), and JavaScript (for dynamic web content) skills to extract and process data if needed.

Q23: Can I use ChatGPT to solve project questions and just copy-paste the answers?

A23: You can use ChatGPT as a tool. In fact, we encourage using LLMs. However, there’s a difference between “learning” and “copy-pasting.” If you copy-paste without understanding, you’re only cheating yourself. You might get the marks, but you’ll miss out on valuable learning that’s crucial for real-world scenarios and technical interviews. It’s better to use ChatGPT to understand the code, debug errors, and learn how to solve problems programmatically.

Q24: What is the purpose of the instructor asking students to refresh questions each time?

A24: The instructor’s goal is to ensure students complete all questions in one sitting. This setup encourages focused work and efficient problem-solving. While this might be challenging for some, it helps develop planning and execution skills. Feedback regarding auto-save features or longer completion windows is being considered for future terms.

Q25: What is JSON, and why is it important?

A25: JSON (JavaScript Object Notation) is a lightweight data-interchange format. It’s easy for humans to read and write and easy for machines to parse and generate. You can ask ChatGPT “What is JSON?” to get a detailed explanation and code examples. Understanding JSON is crucial for working with data extracted from web APIs and other sources.

Q26: What is the overall purpose of this Tools in Data Science (TDS) course?

A26: The purpose of TDS is to equip you with practical tools and skills for data science. It’s a hands-on course, focusing less on theory and more on practical application. The goal is to make you job-ready by teaching you how to collect, clean, analyze, visualize, and tell stories with data, as well as how to deploy your solutions. This course provides you with the foundational toolkit necessary to thrive in the data science industry.Here’s an FAQ based on the provided TDS live tutorial:

Q1: How do I know if my peer review is an outlier?

A1: You won’t know this directly. When you submit your review, the system automatically checks it against rubrics and other reviews. If your score is significantly different from the others for the same submission (e.g., you give a 0 when others give 80-90, or vice versa), it will be flagged as an outlier.

Q2: Will the bonus marks for peer reviews be calculated automatically?

A2: Yes, the system calculates these bonus marks at the backend. If none of your reviews are flagged as outliers, you’ll receive one bonus mark.

Q3: An email mentioned a third project involving an LLM app. Is that correct?

Q3: There are only two main projects for this term (Project 1 and Project 2). The LLM app is actually an opportunity to earn bonus marks, not a full project. You can find more details about this on the Discourse forum.

Q4: How valuable are these bonus marks, and how do they relate to graded assignments (GAs)?

A4: Bonus marks are very valuable. To give you an idea, earning 5 bonus marks is equivalent to getting 5 graded assignments right. This can significantly impact your final grade, especially for difficult parts of the course like ROE. Don’t underestimate their worth!

Q5: Will there be mock ROEs, or will the pattern be similar to past terms?

A5: There isn’t a traditional “mock” ROE. You should look at past ROEs as examples. However, the pattern of ROE questions is never the same from term to term; they are often wildly different. The key is to master the core skills, not a specific pattern.

Q6: What are the core skills I need to master for ROE?

A6: For ROE, you need to be very comfortable with:

• Scraping HTML and PDFs.
• Extracting data from various sources, including databases.
• Using libraries like Pandas and Nominatim.
• Understanding and using CSS selectors (this is crucial!).
• Dealing with authentication when accessing data.
• These skills are often covered in the “optional” videos in Week 2, which you should not ignore, as they are fundamental for ROE.

Q7: How many modules are there in this course?

A7: You can find the full week-by-week breakdown on the course introduction page. It details exactly what will be covered each week.

Q8: Can you give a quick demo of how the LLM app bonus works?

A8: (Instructor demonstrates prompting ChatGPT to create a Python calculator app.) The idea is to use an LLM (like ChatGPT) to generate an application. For the bonus, you’ll:

1. Prompt: Ask the LLM to create an app (e.g., “Create a calculator app in Python”).
2. Copy Code: Take the code generated by the LLM.
3. Run & Debug: Paste it into a development environment (like VS Code), run it, and if there are errors, feed the error message back into the LLM to get it to fix the code. You are not allowed to manually edit the code yourself.
4. Deploy: Get the LLM to provide instructions for deploying the app (e.g., using Streamlit, GitHub Pages, Heroku, Netlify, or Vercel).
5. Record: Record your screen demonstrating the entire process (prompting, running, debugging with LLM, deploying) in a single, unedited 10-minute video. This shows how LLMs can automate tasks like creating apps or cleaning data.

Q9: Where do I upload or showcase my LLM bonus project?

A9: You need to post it in the specific Discourse thread dedicated to the LLM bonus. You can host your app on platforms like GitHub Pages, Heroku, Netlify, or Vercel. If you can’t host it live, you can upload it as a zip file, but this might result in slightly lower marks. Read the pinned Discourse post carefully for all instructions.

Q10: Are there any specific video requirements for the LLM bonus submission?

A10: Yes, your screen recording must be a single, unedited 10-minute video. No editing, trimming, enhancing, or annotating the video is allowed. You should record yourself creating the entire app from start to finish, using only the LLM.

Q11: What is the deadline for the LLM bonus project?

A11: The deadline for the LLM bonus is October 7th. You’ll find this date specified in the Discourse post.

Q12: Is a simple web calculator considered an acceptable app for the LLM bonus?

A12: The criterion is “how impressive the app is.” While a web calculator might be simple, the goal is to see your ingenuity. The project is open-ended, so you decide what kind of app you want to create using the LLM. It should be interactive and accessible via a browser.

Q13: I have no prior programming experience. How can I learn everything quickly enough?

A13: The key is “time management” and “being honest with yourself.” While it might seem like a lot, the course provides all the learning resources. It’s about being proactive. Read all the provided materials, including the “optional” sections and hyperlinks, as they contain valuable information. If you’re new, you’ll need to dedicate more time to practice.

Q14: What kind of data and tools will be used in the projects?

A14: Projects typically involve large, unclean datasets. You’ll be given a set of questions (around 10-15) to answer by processing this data. This often involves using a mix of the tools covered in the course (e.g., HTML/PDF scraping, data extraction from databases, data cleaning, visualization). Sometimes, the data might come in various formats (e.g., Parquet files, SQL databases, PDFs), requiring you to combine information from multiple sources.

Q15: What kind of questions can I expect in ROE?

A15: ROE questions are highly varied and unpredictable; they are “completely different beasts” each term. You cannot simply eyeball an answer. They test your core skills and intuition. For instance, past ROEs have involved:

• Scraping ball-by-ball cricket commentary (including handling programmatic elements that required scrolling to reveal data).
• Extracting and cleaning data from deliberately unclean PDFs.
• Combining information from three different data sources (Parquet, SQL, PDF) to answer complex questions. The key is to master the core skills and be prepared for anything.

Q16: Is ROE open book?

A16: Yes, ROE is completely open book and not proctored. You can use any resources, including ChatGPT, friends, or the internet. However, this highlights that the exam tests your intuition and ability to quickly apply your knowledge, not just memorization. You won’t have time to search for every answer; you’ll need to know the core skills well enough to execute efficiently.

Q17: I don’t have Excel. Do I need to purchase it for the course?

A17: No, you don’t need to purchase Excel. You can use the free cloud version of Microsoft 365, which is sufficient for most tasks. The course also introduces powerful 3D visualization tools within Excel that can create impressive visuals.

Q18: Will JavaScript/CSS skills be required for the projects or ROE?

A18: Yes, especially for ROE. While you’re not expected to be an expert developer, you must be very comfortable with CSS selectors. Week 2 covers this. Understanding CSS selectors is critical for web scraping and extracting specific data elements from web pages.

Q19: I’m facing an issue with GA0, where the modified date of the files changes, causing errors. How can I fix this?

A19: This is a known Windows-specific issue that occurs when using the built-in “extract all” feature for zip files. To resolve this, use the unzip command in Git Bash instead of Windows’ default extractor. This should preserve the original timestamps.

Q20: What is the purpose of refreshing GA0 questions every time I open it?

A20: The instructor’s objective for GA0 (and GA1) is for you to complete all questions in one sitting. This setup encourages focused work and efficient problem-solving. While this might be challenging for some, it helps develop planning and execution skills. Feedback regarding auto-save features or longer completion windows is being considered for future terms.

Q21: What kind of questions will be in Project 1?

A21: Project 1 typically involves a large, potentially unclean dataset. You’ll be given a set of 10-15 questions to answer by processing and analyzing this data. You’ll need to create charts and present your findings. Depending on the dataset, you might need to use skills like HTML/CSS/JavaScript for data extraction.

Q22: Will the projects require HTML, CSS, or JavaScript?

A22: It depends on the specific dataset provided for the project. Last term, the dataset required web scraping from an HTML source, including dealing with deliberately “noisy” HTML. So, yes, you should be prepared to use your HTML, CSS (for selectors), and JavaScript (for dynamic web content) skills to extract and process data if needed.

Q23: Can I use ChatGPT to solve project questions and just copy-paste the answers?

A23: You can use ChatGPT as a tool. In fact, we encourage using LLMs. However, there’s a difference between “learning” and “copy-pasting.” If you copy-paste without understanding, you’re only cheating yourself. You might get the marks, but you’ll miss out on valuable learning that’s crucial for real-world scenarios and technical interviews. It’s better to use ChatGPT to understand the code, debug errors, and learn how to solve problems programmatically.

Q24: What is the purpose of the instructor asking students to refresh questions each time?

A24: The instructor’s goal is to ensure students complete all questions in one sitting. This setup encourages focused work and efficient problem-solving. While this might be challenging for some, it helps develop planning and execution skills. Feedback regarding auto-save features or longer completion windows is being considered for future terms.

Q25: What is JSON, and why is it important?

A25: JSON (JavaScript Object Notation) is a lightweight data-interchange format. It’s easy for humans to read and write and easy for machines to parse and generate. You can ask ChatGPT “What is JSON?” to get a detailed explanation and code examples. Understanding JSON is crucial for working with data extracted from web APIs and other sources.

Q26: What is the overall purpose of this Tools in Data Science (TDS) course?

A26: The purpose of TDS is to equip you with practical tools and skills for data science. It’s a hands-on course, focusing less on theory and more on practical application. The goal is to make you job-ready by teaching you how to collect, clean, analyze, visualize, and tell stories with data, as well as how to deploy your solutions. This course provides you with the foundational toolkit necessary to thrive in the data science industry.