Home/ Questions/Q 8191483
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-06-07T03:55:49+00:00

I’m looking at creating a heatmap of numerical data spread over various locations within

  • 0

I’m looking at creating a heatmap of numerical data spread over various locations within a building. I’ve spent a few hours researching data mapping, etc. and am looking for some advice. I am new to GIS. The majority of options available are mostly tile APIs that use lat/long, and are overkill for my requirements…

Ultimately, I just want to output a background image (a floor plan) with the heatmap overlay demonstrating areas of high intensity. The data is bound to specific locations (example, activity level: 14, location: reception entrance) and so is not randomly distributed over the map. Data has timestamps, and the final objective is to print PNGs of hourly activity for animation.

I feel like I have two options.

I like this tutorial (http://dylanvester.com/post/Creating-Heat-Maps-with-NET-20-%28C-Sharp%29.aspx) as it offers a huge amount of flexibility and the final imagery is very similar to what I would like – it’s a great head start. That said, I’d need to assign locations such as “reception entrance” to an x,y co-ordinate, or even a number of x,y co-ordinates. I’d then need to process a matrix prior to every heatmap, taking data from my CSV files and placing activity values in the appropriate co-ordinates.

The other option I think I have is to create a custom shapefile (?) from the floor plan. That is, create a vector graphic with defined regions, where each is attributable to a taggable location. This seems the most flexible option, but I’m really, really struggling to find out how to create shapefiles?

My unfamiliarity with GIS terminology is making searches difficult. The latter seems the most sensible solution (use the shapefile with something like https://gist.github.com/1370472) to change the activity values over time.

Links found:

  • guthcad.com/cad2shape.htm (but don’t have CAD drawing, just raster floorplan)
  • stackoverflow.com/questions/4014072/arcgis-flex-overlay-floor-plan-png (unhelpful, don’t want tiled)
  • oliverobrien.co.uk/2010/01/simple-choropleth-maps-in-quantum-gis/
  • gis.stackexchange.com/questions/20901/using-gis-for-interactive-floor-plan (looks great)

To summarise: I’d like to map data bound to locations within a building. There’s very good code in a C# tutorial I’d like to use, but the linking of activity data to co-ordinates is potentially messy (although could allow for describing transitions of activity between locations as vectors between co-ordinates could be used…). The other option is to create an image with regions that can be linked to CSV data by something like QGIS. Could people with more experience suggest the best direction, or even alternatives?

Thank you!

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    I recently did something similar for a heatmap of certain events in the USA.

    My input data was simply a CSV file with three columns: Latitude, Longitude, and Number of Events.

    After examining available options, I ended up using GHeat.Net. It was quite easy to use and required only a little modification to meet my needs.

    The output is a transparent PNG that I then overlaid onto Google Maps.

    Although your scale is quite different, I imagine the same solution should work in your case.

    UPDATE

    If the x,y values are integers in a reasonably small range, and if you have enough samples, you might simply create a (sparse?) array, with each array element’s value being the number of samples at that coordinate. Identify the “hottest” array element (the one with the most samples) and equate that to “white” in a heat map, with lesser values corresponding to colder colors (or in other words, normalize all values in the array using the highest value and map the normalized values to a color scale). Map the array to a PNG.

    Heat maps like GHeat create a sphere of influence around each data point. Depending on your data, you may not need that.

    If your sample rate is not high enough, you could lift the sphere of influence code out of GHeat and apply it to your own array.

    The sphere of influence stuff basically adds a value of “1” to the specific coordinate in the data sample, and also adds a smaller value to adjacent pixels in the map in order to provide for smoother-looking maps. I don’t know the specific algorithm used in GHeat, but the basic idea is to add to the specific x,y value as well as neighbors using a pattern something like this:

    0.25 | 0.5 | 0.25
-----------------
 0.5 | 1.0 | 0.5
-----------------
0.25 | 0.5 | 0.25
    • 0