Overview

Part 1: Approach and Methodology

 

API Setup

​

An API (Application Programming Interface) is like a translator that lets two different software programs talk to each other and share information. It helps apps or systems work together by following a set of rules. Using API is like letting different software communicate in their own language.

For that, I created accounts on OpenAI and Reddit to access their respective APIs. These APIs allowed my bot to retrieve data (e.g., comments) from specific subreddits and use OpenAI’s language processing capabilities for summarization and analysis of information. Following is the summary of the steps to do so:

Creating secret key for openai:

• Make an open ai account.

• Search for API keys in the platform

• Create new secret key

 

​

​

​

​

​

​

​

​

​

​

 

 

 

 

Creating secret key for reddit:

• Make a reddit account.

• Open reddit preferences

• Create a secret key

 

​

​

​

​

​

 

 

 

 

 

 

 

Installing Libraries

​

A library is like a collection of pre-written code that helps your programs do specific tasks without having to write all the details yourself. It’s like a window through which you can talk to your computer. In our case, I needed to install libraries of openai and reddit in my mac. For that I followed these steps:

• Open terminal

• Type-in these commands: 

pip3 install openai (openai: For natural language processing and summarization)

pip3 install praw (To interact with Reddit’s API and extract discussion threads)

​

Coding and Execution

​

I have very limited knowledge of coding, so for the coding part I prompted chatgpt with this command:

Give me a script for python to:          

1. Search Reddit for threads on a specific topic.

2. Retrieve thread data (like title and text).

3.  Use OpenAI to analyze the content.

​

After I copied my code in visual studio editor app. This app is like a notebook facilitating script editing. To enhance the codes, I needed to modify a few details such as:

In the fetch_reddit_threads function, I set the limit to 20, it defines how many subs to retrieve each time. If the limit is too big, the bot will be less precise.

 

​

​

​

​

​

​

​

I also modified the temperature. This is a very cool option. The temperature varies from 0-2. If the temperature is set to 0, openai only used the retrieved data from the source and only uses them to analysis the data, yet as much as temperature is increased, openai can be creative using other data to analyze and produce answers, yet the accuracy of the responses may reduce by that. I decided to set it to 0.7 for this project.

​

​

 

​

​

​

​

 

 

​Output

​

For the last step, you need to run the code in python. By doing so, the bot provided concise, structured insights based on discussion threads without requiring me to manually read through dozens (or even hundreds) of posts, comments and discussion!

For example I search “educational technology” in Concordia subreddit. For that I can see a summary and analysis of what Concordia students in this group say about my master’s program. Here is the link to the result I received within seconds.

​

​

 

Part 2: Lessons Learned

 

This experiment highlighted several key insights:

• AI as a Time-Saving Tool: The bot significantly reduced the time needed to gather and analyze discussions, proving its value as a productivity enhancer for instructional designers.

• Limitations of Automation: While efficient, the bot’s ability to contextualize complex topics was limited. Manual review of some outputs was still necessary.

• Future Potential: This method has the potential to become a valuable skill in my field, as AI-driven tools can assist in rapidly synthesizing user feedback and staying informed on industry trends. Yet, this bot is very primary and basic and runs only in python. To make it look like an app with a more user-friendly interface, there is a need for advance programming skills that is beyond my current skills.