select data
Here, we can examine the growth of covid cases and deaths at a national or international scale
Data pulled from Johns Hopkins Github
This plotly line graph displays Covid-19 data at timeday and time day+14
14 days is the average incubation period for coronavirus as described by the Centers for Disease Control and Prevention
This chart illustrates not only the growth in total cases, but also the change in daily average
For interactivity, click the chart to visit TableauPublic
To the left, a treemap of words representing the top tf-idf values in tweets on days where the acceleration of Covid-19 cases in Richmond, VA does not increase at 14 days
For interactivity, click the treemap to visit TableauPublic
Initially, the RandomForest model achieved a score of 82.6% by utilizing the top 5000 tf-idfs as features over 91 days of treets. This is with the model predicting an increase in daily_avg 100% of the time However, by simply looking at analysis of the Department of Health data, you can see that 80.2% of the days in our data set contributed to an increase in daily average, in the first place. Manually lowering the probability threshold to .25 for 'decrease' results in slighlty more resolution in our confusion matrix.
Click here for slide deck