SentiShelter

AI chat
Logo
Website ss

Inspiration

Have you ever thought about buying a house and found yourself considering the climate of the region? What about how high your rent might be this month, or the next natural disaster, just looming around the corner? We all have been there. Now, imagine having a tool that not only provides insights into the shifting perspectives about climate change, but also about its relationship with the housing market.

What it does

The goal is to utilize sentiment analysis to track people's feelings, in their discussions concerning the interplay of climate change and housing as time progresses. 1) Sentiment Analysis for the Insights

Using Reddit data about climate change and housing, we can find answers to the questions of 'why' and 'how' climate change is discussed in Reddit comments and delve into the root causes of this conversation. 2) Website Built for Users
To communicate these powerful insights, we built a website where users can view these sentiments
Bar charts and scatter plots help explain relation between different independent and dependent variables.
We implemented a chatbot feature using LangFlow, allowing users to ask their most questions with a real-time response.
Chatbot is tailored to address topics related to climate change and the housing market.

How we built it

We used the Kaggle dataset provided at: https://www.kaggle.com/datasets/pavellexyr/the-reddit-climate-change-dataset/data
Using a Kaggle notebook, we were able to clean, reduce, and analyze our data using python. -We employed Spacy's natural language processing techniques to teach our code how to read, assess, understand, manipulate, and make sense of human language. Using this approach, we identified key terms associated with climate and housing to successfully monitor the frequent mentions of people's names, organizations, or locations in Reddit all comments.
For our project to deliver a data driven storytelling initiatives, we used python packages like seaborn, matplotlib to create visually appealing graphs.

Challenges we ran into

-The dataset takes huge memory storage so it took time to load. We took samples of data that would not skew the insights. -Figuring out with type of tools needed time and discussion as well.

Accomplishments that we're proud of

We were proud of what was already accomplished with SentiShelter. However, there is always more to be done to make the analysis even more accurate.

What we learned

-We learned about AI not just using AI to analyze events but also implementing it in our website to answer our findings.

What's next for SentiShelter

-Next we can assist companies with other problems as well. We will conduct monthly sentiment analysis from January 2020 to August 2022, focusing on this subpopulation. Our goal is to identify any differences in trends compared to the initial analysis. Furthermore, we are dedicated to creating our own sentiment analysis tool using unique test data to ensure the highest quality insights for our users. We're enthusiastic about enhancing our NLP capabilities on a global scale. We're thrilled about the potential collaborations with climate experts and housing companies to demonstrate SentiShelter's real-world impact, ultimately simplifying the housing market for all. We can't wait to bring this simplification within reach of millions.