Project Luther - Web Scraping and Linear Regression
Projects | | Links: Related Posts
A journey through web scraping to accumulate data for linear regression.
Exploring the impact of various features related to property tax and valuation on neighborhood stability.
Back Story
An email soliciting assistance in exploration and investigation:
Joseph
Greetings. I was very intrigued in our recent discussion about Data Science. It was interesting to learn about how folk use computing to analyze data to draw insights.
Well, the group I work with would like your assistance to chase a curiosity. Is there a way to discern from publicly available housing data what factors influcence neighborhood stability in the sense that people stay in their houses for longer periods of time?
thanks in advance,
Frank Russell Trated
Initial response…
Frank,
I’m including here the presentation we reviewed which described my initial assessment of the possiblity of using publically available data provided by the local governments for Property Tax purposes.
I will continue to explore this topic and provide a final report after further investigation and analysis.
regards,
Joseph
Final response…
Frank,
It was a pleasure to be able to explore this topic.
However, as suggested during the MVP review, the data doesn’t seem to suggest we create models via linear regression here with significant predictive power.
I was indeed able to scrape far more data. But an analysis of the modeling demonstrated the problem here isn’t with the amount of data. More data isn’t going to (and didn’t) create a signal to capture where one may not exist.
I am providing links below to the final presentation and related items.
regards,
Joseph Hamilton