Density Heatmaps

Converting observation points to a 3d-heatmap

Often the data available is in form of points. For example, you may have an excel-file with many rows each having a timestamp and some observed value - which can be interpreted as (time,value) points. Observations may be from some process you are monitoring (e.g. physical or financial process) and indicate process activity or quality. In another scenario, you have location information as (x,y)- or (latitude,longitude)-coordinate wherein each point indicates that some interesting event has happened on that location.

Visualization is a great way to summarize of the data and get a quick overview on what's going on. In this case, you probably are interested to know which areas have the most points and idenfity clusters or trends. When the number of points is low, plotting each point into a graph or on a map works fine. But when the number of points increases, scatter plots become a mess. Then you need another way to analyze point density.

HeatMiner® provides an easy way to convert point-based data into colorful heatmaps. For example, when given this file as the input, HeatMiner® can generate the following 3-dimensional heatmap. The height and colors indicates the density of the given input points. Visualization is based on Java3D, so in case you don't see the 3D-heatmap, your web browser probably does not support Java3D. In that case, check the Snapshots.

Location probability distribution

You can rotate, move and zoom the heatmap image with your mouse. Hold left button down to rotate the image and use the right button to move it. Left button with Alt-key pressed will zoom in and out. Click image title to open a larger view.

For comparison, here is the same data plotted with gnuplot (which is a fabulous research tool, but reaches its limits with this many points). You cannot really tell from that mess which area has the most points, can you?

Try HeatMiner!

Some of these great HeatMiner® heatmap visualizations are now available as-a-Service at the Cloud'N'Sci.fi algorithm marketplace. Go to the HeatMiner homepage and create heatmaps from your own data using free evaluation campaigns! New heatmap types and demos are introduced frequently at the new HeatMiner wiki pages so keep an eye on it for updates.

Want to see more?

Check the latest HeatMiner solutions and demos!

<< Variance Heatmaps | Good-old Demos | 4D-heatmaps >>

Like it?

Become a Facebook fan!