Live Session - 2#
Duration: 0h 29m
Here’s an FAQ-style summary of the Tools in Data Science (TDS) live tutorial:
Tools in Data Science (TDS) Live Tutorial FAQ#
Q1: I’m having trouble hearing; your voice is buzzing and echoing.
A1: Apologies for the audio issue; this is the first time we’ve encountered this. I’ve adjusted my microphone, and hopefully, it’s clearer now. Please let me know if it’s still buzzing.
Q2: Was the recording for the previous session available, as I couldn’t find it on YouTube or Drive?
A2: Unfortunately, the recording for the last session is not available. There was a technical issue during the recording, and we don’t have it.
Q3: Could you please summarize the content of the last session in a few minutes?
A3: The last session was a basic introduction to the course and its content, which you likely already covered in your course orientation. I went over the different topics planned for each week. We discussed the various tools you’ll be introduced to, primarily Python and Excel. Additionally, you’ll learn about visualization tools like Tableau, networking tools such as Kumu, geospatial analysis with QGIS, and narration tools. The main goal isn’t for you to become an expert in every tool, but to understand how to identify a problem, select and use an appropriate tool to solve it, and then effectively communicate your findings.
Q4: Will there be additional sessions, and how can I get help with doubts?
A4: Yes, we are planning supplementary TA (Teaching Assistant) sessions every week, covering different topics. We’ll get in touch with you about the schedule soon. If you encounter difficulties with any particular topic, you can inform the TA, and they can conduct sessions to address those issues. Beyond that, I encourage you to form student groups to help each other, and please make full use of Discord. You can tag instructors there if you have specific questions, and we’ll respond as quickly as possible.
Q5: I found an discrepancy in Graded Assignment 1, Question 1. My Excel calculation for “banana green” in Malappuram resulted in 3350, but the system shows 3650 as correct. Can you verify this?
A5: Thank you for bringing this up. It seems several students encountered a similar issue. I’ll need to investigate this with the team. Could you please raise this specific query on Discord and tag me (@Dixon)? That will ensure I see it directly and can get back to you with a clarification as soon as possible.
Q6: Can you explain the difference between categorical ordered, categorical unordered, and composite data types, especially as they relate to questions in Graded Assignment 1 (specifically Questions 5 and 7)?
A6: While I haven’t reviewed the exact questions from the assignment yet, I can give you a general overview.
Categorical Ordered vs. Categorical Unordered: This distinction primarily depends on the meaning of the variable, not just its structure (e.g., whether it has numbers or not). If there’s an inherent, meaningful order or hierarchy among the categories, it’s categorical ordered. For instance, “strongly agree,” “agree,” “neutral,” “disagree” has an order. If there’s no such intrinsic order (like colors: red, blue, green), it’s categorical unordered. The presence of numbers or dots in the variable name (e.g., ID_1, ID_2) doesn’t automatically make it ordered; it’s about whether the values themselves imply a rank or sequence.
Composite Data: A composite variable combines multiple pieces of distinct information into a single data point. The classic example is an IP address (e.g., 10.108.8). Although it might look like a single number, each segment (10, 108, 8) carries individual meaning (e.g., continent, region, sub-region) that, when combined, forms a complete identifier. Similarly, if a column combines, for example, a country code and a district code (e.g., “FR.AR”), it’s likely composite because “FR” has one meaning, and “AR” (in conjunction with FR) has another. The key is that these parts could stand alone as meaningful attributes or together represent a more complex, structured entity.
For Question 5, which included entries like “FRA.1_1”, while it looked structured, the “1_1” part likely functions as a mere serial number without deeper meaning or order relative to other numbers, making it categorical unordered if the “FRA” is the primary category. However, for Question 7, with entries like “FR.AR”, where “FR” could denote France and “AR” a specific region, it would be considered composite due to the combination of distinct, meaningful information.
Q7: I’m still unclear about why “FRA.1_1” in Question 5 is not composite, but something like “FR.AR” in Question 7 is.
A7: Let’s clarify that. For “FRA.1_1,” the “FRA” part represents a category (like a country), but the “1_1” part itself is typically just a serial number or an identifier without additional distinct meaning that contributes to a hierarchical composite structure beyond just its numeric value. It’s essentially “Category A, serial number 1.” The “1_1” doesn’t inherently contain separate, meaningful sub-components that, when combined, create a new, richer piece of information like an IP address.
However, in the case of “FR.AR” from Question 7, “FR” clearly signifies a country (France), and “AR” likely represents a distinct administrative division within that country (like a county or district). Here, “FR” and “AR” are two separate, meaningful pieces of information that are combined to form a more specific location identifier. This combination of distinct, independently meaningful attributes is what makes it a composite data type. The dot often serves as an indicator that different meaningful segments are being joined.
Q8: What will be the difficulty level of the course weeks, especially since I’m also managing another intensive subject like MLP?
A8: The difficulty level can be subjective and depends on your prior experience. Since you don’t have direct experience in data science or using Python for data science problems, I’d say the course might start feeling more intensive for you around Week 4. This is when we begin to delve into more complex topics like data modeling and other advanced concepts.
Compared to subjects like MLP (Machine Learning and Programming), this course is generally considered easier because it’s purely application-oriented. We won’t be focusing heavily on the underlying theory or complex mathematical derivations. Instead, the emphasis is on applying tools and techniques to solve real-world data problems. So, while Week 4 onwards might require more focused effort, the overall approach is practical and less theoretical than some other courses.