Interactive modeling allows for calculating permutations from factor values to find the key groupings. AppUp can model your mixed-effects data to better structure an analysis.
Analysis allows for creating forecasts and predictions as well as determine what factors have importance and what should be visualized.
Visualizing the data helps communicate relationships, importance, patterns, and more. AppUp can help you tell the story your data wants to tell.
This visualization was put together to show popular starting points and end points across the TwinCities via the NiceRide bike share. Each trip in 2018 was plot with a line between stations. The darker areas highlight the more popular areas. Visualizing the data in this way also reveals where additional stations may benefit the city.
In order for me to create this visualization I had to merge existing technology. The data was collected by scraping the New York Times election results page and the 2D plot election results plot was rendered in ggplot. The county densities used for elevation were scraped from Wikipedia, the 2d elevation plot was also made in ggplot. For the interactive 3D portion, Rayshader which leverages rgl, was used.
This plot was made by pulling boundary information for counties in the midwest, along with population data from the Census Bureau.
The plot shows that within the midwest, most of the population lives in or near urban areas. Even by dropping cities such as Springfield, IL or Bloomington, IL the population in midwestern cities surpasses the rural population. Postal code level data would provide greater insight.
Created a map of destinations by start station. Points sized by destination frequency from the selected station station. Popup includes number of rides and the option to change selected start station.
Heatmap was created in D3, while the single image exports were created in Tableau. The stations were clustered in R.
Data visualization shows the population change in percent from 2010-2016. Color bins were manually selected and the shape files were provided by the MN DNR.
There was a lot of data to link in this analysis. I started by grouping similar subreddits and used Tableau for my initial exploratory analysis. For the author's age I wanted to show the author's age at the time of submission, so I made a calculated field that returned the difference in the date of the author and the date of the submission.
Comparing political subreddits needed context, which is why the grouping of all other subreddits and subreddits that commonly have throwaway accounts were needed.
Created this from the 2011-2013 ODOT data because someone was complaining about cyclists not paying the gas tax where they lived (Portland). Taking into account the damage (1/9,600 the damage of a car, and roads cost a pretty penny to repair) and that many cyclists have cars if not at least a license, (89%) it became clear that cyclists have been subsidizing motorists in Portland. [View Larger]
Data comes from the 2012 Minneapolis ALPR data inside the date range August 30th to November 29th. For readability's sake any coordinates with fewer than 5 pings were removed. Circumferences increase starting at 51 pings, red is 100 pings, and overlaps are from coordinates very close to each other. The overlap is prevalent in the bottom right with the values 100+80+76+54+9 = 319. I plan on doing a traditional heatmap as well using Google Maps to interactively explore if there is enough interest. [View Larger]
Yahoo had just acquired Geocities, Netscape Navigator was all of the rage, and notepad.exe was my editor of choice.
With a few years under my belt I was finally able to start freelancing my web development skills.
So much Flash; so much incompatibility.
I joined an amazing firm in 2004. The firm is now defunct, but their sister company Solid State Networks is still alive and kicking.
Working with them taught me a lot. I was introduced to large scale databases, SEO, and the value of internal build tools.
After establishing myself as a Minneapolis freelancer for over a decade I decided it was time to operate under a business name. I was releasing web applications using dokku with frquency, and decided on AppUp because of thecommand.
For the most part, as AppUp, I work with smaller companies providing insights they might otherwise be unable to find.