Omdena Chapter Page: Texas

Omdena Texas Chapter - Omdena Chapters

 

Project Started: September 22
Duration: 4 Weeks
All Data Science Skills Welcome!

Applying Data Analytics & ML methods to improve Homelessness

The Background

According to recent statistics, there were 580,466 people experiencing homelessness in America. Most were individuals (70%), and the rest were people living in families with children. They lived in every state and territory, and they reflected the diversity of our country. This is a big societal issue that many city councils and local governments are trying to solve.

The Problem

Below some important socio-economic facts about people experiencing homelessness. 

1. Both nationally and in Texas, the inability to afford housing is one of the leading causes of homelessness. A lack of investment in affordable housing options combined with stagnant wages leaves stable housing out of reach for many. In Austin, a minimum wage worker would need to work 125 hours a week just to afford a one-bedroom apartment!

2. People experiencing homelessness are likely to be single adults with no easy means to make a livelihood. This can be improved with the right policies.

3. Ending homelessness does not mean individuals and families will never again experience homelessness. Instead, it means that as a community we will have a systematic response that can address immediate needs, quickly connect people to housing, and provide services to ensure long-term stability.

The Project Goals

1. Analyze datasets and perform EDA on homelessness in different regions

         1. Supply demand (NLP and which NGO to reach out to? (can be scraped)

         2. Dashboard Report localized on USA/Texas COC. Eg.State of Homelessness: 2021 Edition – National Alliance

2. Build prediction ML models for resource allocation purposes something like this to predict and help those at risk of    eviction/homelessness (like Chronic Homelessness Artificial Intelligence model (CHAI)) 

        1. Predictive modelling on bed occupancy through seasonal forecasting 

3. Create a WebApp tool and data visualization methodology (using GIS, Streamlit, Matplotlib and come up with meaningful proposal

         1. One stop affordable housing webApp for NGOs based on different features (eg. credit history, demographic parameters)

         2. GIS visualisation of where is what concentrated: Homelessness pockets vs resources (if possible into different subcategories)

The Tasks & Timeline

Week 1

(Data roundup)

Week 2

(EDA)

Week 3

(Data Vis & ML)

Week 4

(WebApp)

– Data collection of homelessness across the USA/Texas

– Data  Preprocessing

– GIS info on campsites (Images idea to be explored)

– Exploratory Data Analysis (EDA)

– Correlation with housing, income, health and health

– GIS analysis

–  Build prediction models

– Start building WebApp for matching homeless people and available  resources

– Finish building WebApp tool

– Write article

– Plan for final presentation on Nov 4

 

Texas Chapter Leads

Kaushik Valluri

A technologist with over a decade of experience in product and business development for startups and large corporates. Most recently, he designed and developed a Voice/NLP product for the healthcare industry. In his spare time, he coaches people who stutter.

 

 

 

 

 

The Learning Outcomes

1. Determine the correlation between homelessness and housing, employment, income, and health

2. Learn to build predictive models and other ML techniques

3. Build a webApp tool