Difference between revisions of "User:KWeathington"
From REU@MU
KWeathington (Talk | contribs) |
KWeathington (Talk | contribs) |
||
| Line 39: | Line 39: | ||
*Began developing theory and equation for a "wastage index" | *Began developing theory and equation for a "wastage index" | ||
| + | |||
| + | <h2>Week Six</h2> | ||
| + | |||
| + | *Implemented a K Nearest Neighbors plot to help identify optimal eps given a minPT value on certain subsections of the data | ||
| + | |||
| + | *Added hulls around individual clusters to make more user friendly | ||
Revision as of 19:41, 6 July 2018
Mentor: Dr. Shion Guha
Lab Github: https://github.com/marquettecomputationalsocialscience
Personal Github:https://github.com/KatyWeathington
Week One
- Orientation
- Data discussion meeting and updating new team members on previous work and lab goals
- Background reading on various spatial clustering algorithms and use of data science concepts in police departments
Week Two
- Compiled a preliminary list of clustering algorithms sorted into family, with a note of potential sources of biases and a summary of the steps
- Demonstrated a variety of clustering algorithms on random data
- Explored Kernel Density Estimates
Week Three
- Expanding a dictionary for sanitizing address data
- RCR Training
- Made a poster for upcoming Northwestern Mutual event
- Opened API keys
Week Four
- Added basemap to DBScan visualization
- Created several DBScan clustering models of crime in Milwaukee
- Presented research at Northwestern Mutual technology showcase at announcement of the Data Science Institute
Week Five
- Reporting halfway progress
- More DBScan and nearest neighbors based calculation of epsilon
- Began developing theory and equation for a "wastage index"
Week Six
- Implemented a K Nearest Neighbors plot to help identify optimal eps given a minPT value on certain subsections of the data
- Added hulls around individual clusters to make more user friendly
