@hayleefay . You are in charge of this visualization. We want to use Bokeh and the surf3Dtime directory to make a 3D surface plot of one of the household distributions (c_{j,s,t}, n_{j,s,t}, b_{j,s+1,t+1}) with a line plot of one of the aggregate time paths (r_t, w_t, K_t, L_t, Y_t, C_t) and a slider bar for the time period. See the pencil drawing example below.

We want:

1. The top 3D plot to be manipulable in terms of the view (perspective) and zoom.
2. Buttons on the side that allow the 3D plot to be changed among c_{j,s,t}, n_{j,s,t}, and b_{j,s+1,t+1}.
3. Buttons on the side that allow the aggregate time path to be changed among r_t, w_t, K_t, L_t, Y_t, and C_t.
4. A slider bar for time period that changes the highlighted point in the lower-pane aggregate time path plot and the distribution in the upper 3D plot.

@DoraSzasz has said she is happy to help you at any stage with this visualization.

@hayleefay. I think we can be done with the 3D surface plot after making the following two changes.

1. Get the vertical axis label in the 3D surface upper-div plot to be vertically oriented.
2. Make the lower-div plot's buttons vertically oriented instead of horizontally oriented so they take up less space.

@DoraSzasz @hayleefay. This issue is to start the development of a visualization of marginal tax rate and effective tax rate data. In particular, we want:

1. Dynamic visualization using the Bokeh library
2. 3D scatterplots of tax rates of individuals as a function of capital income and labor income
3. 3 separate plots, selected by buttons: marginal tax rate of labor income, marginal tax rate of capital income, effective tax rate
4. Slider bar for age
5. Slider bar for year, current year through current year + 9

An example of from one of our papers is the following.

Below is a whiteboard drawing of the visualization that we want to make.

@hayleefay. The 3D surface plot with aggregate variable time path below and time slider bar in /surf3Dtime/ looks really good. Let's do a 2D version of that image where the upper-div is a 2D plot with J number of lines. We can put this in a new folder entitled, /surf2Dtime/. A 2D example of that surface for individual savings b(j,s,t) would be the following.

Note that this will require:

1. A legend for the lines
2. Different markers for the points
3. Different line styles for the lines

Let's use Django to make a web front-end that creates an interactive gallery of data visualizations. And let's useHeroku to host. The hope is that:

1. New visualizations can be easily added.
2. Individual visualization page is fully interactive.
3. Individual visualization page has detailed description.
4. Individual visualization page has link to open source repository for the source code and underlying model.

@hayleefay @DoraSzasz