Housing Team Week 3 Wrap-Up
Project Progress
This week, our team was able to spend time thinking and talking about:
Our takeaways from last week’s client meeting
Project goals
Methods to use when working to reach project goals
New ideas for our project
DataCamp Courses Completed
Web Scraping in R
Intermediate R
Web Scraping
We are currently trying to scrape data from housing assessor websites such as Iowa Assessors (data from Beacon and Vanguard) and Trulia. If we are able to successfully do so, images of houses and other information can be utilized to train our AI model(s).
Kailyn, Gavin and I (Angelina) focused on learning more about web scraping using R through the DataCamp Course called Web Scraping in R. We also began to follow steps for scraping the web using other tutorials.
AI Modeling: Vegetation Model
To get an idea of how to use AI for the project, a sample model was trained to determine if a house has vegetation or no vegetation. About 750 images in total were used to train, validate and test the model.
The model was successfully trained and process did not take long.
Happies
Creation of AI model was quicker than anticipated. We can spend more time into other aspects of the project (ex. data collection).
Kailyn was able to set up her blog!
We now have a better idea of the project’s methods and goals.
Crappies
- Starting the web scraping process. When it comes to web scraping, we need to understand more than what the DataCamp course was teaching.
Some Plans for Next Week
Get a rough database running with housing info. We may be able to utilize a web interface that will allow us to more efficiently filter through houses.
Familiarize ourselves with web scraping.
Meet with client(s) to show the project’s progress.
Finish scraping data for Independence, Slater, New Hampton, Grundy Center, or other cities if needed.
Questions
How many of you have some web scraping experience?