The spatial-temporal cube of ArcGis : 1- discovery

One of the domains of GIS where still remains a lot to be done is the field of temporal analysis  . While almost all the available tools focus on the spatial evolution of a phenomenon, we are left helpless when trying to visualize or analyse a phenomenon that changes as a function of time as well.

For some time, we have the support of animation tools that allow us to see a sequence of maps and perceive the change over time. Regardless of how useful they are for communication, they are useless to perform a serious data analysis. In this series of articles, we will discuss new  ArcMap and ArcGis Pro tools available  in the form of a Toolkit Tools model for the  spatial-temporal exploration (Time Pattern Mining). An example has been selected in order to process, at first with ArcGis Pro, and, later, with ArcMap.

What is a Spatial-Temporal Cube ?

The following image summarizes the answer to this question.


On the left we have our points. They are distributed in the XY space, exactly the distribution we usually study and analyse by using the ArcGis tools. But we have several points located in the same place that correspond to different moments. This is the reason why the points are also distributed along the Z axis, the time axis. The space-time cube will allow us to analyse the distribution of our points on these three axes.

First important thing to remember: The tools that we are going to be using to analyse the spatial-temporal cube are not apt to perform the analysis of a given phenomenon based on its attributesThey are limited to the analysis of the occurrence or non-occurrence of a phenomenon (the point exists or does not exist).

The process is simple; we will determine the size of a box  (bin), such as pixels or raster cells, but in three dimensions. And we will determine the number of points contained in each box. It is this number that we will visualize and analyse.

Let’s use an example in order to follow the process in detail.  We have the weather forecast containing, that includes the gusty winds forecast. These forecasts are for a total of 8 days, every 3 hours, for an area that includes the Brittany.

A look at the attribute table let us see the attribute ” gust “, the speed of the wind gusts in knots, and, another field named Datehour, indicating the date and hour of each forecast.

Maybe one day it will be possible to analyse the distribution of the gust values as a function of time. But for the time being, we are limited to study the occurrence of the points, we can, for example, define a threshold of speed that we consider as risky and analyse the distribution of the points which exceed this threshold value.

If we define the threshold value as 20 knots we will create a new layer of data containing only the points exceeding this value. The points that do not exceed this value will not be present. We do that by selection of attributes, gust> = 20, and we export the selected points.

That’s it; we have now a layer to manufacture our cube.


Creation of the space-time cube

To create the cube, which will be stored under a NetCDF format ( .nc ), we will use the tool   Create a space-time cube Toolbox   Tools models for exploration spatial-temporal .

Once issued the order, the settings window appears.

The   input entities   correspond to our layer of selected gusts ( > 20 knots).

The   space-time cube output   will be our cube.

The   time field   will serve to define the Z axis and it must absolutely be it a Date type field. For those who are always with shapefiles, remember that the date fields can only contain the date but never the time. If you are going to work with time lapses of less than 24 hours (a day) you must have a class of geodatabase entities mandatorily .

For the time being we do not have count with a model cube to recover the others cubes information. If we make a series of cubes, and the first serves  as model , we can be certain that the steps and the location of the boxes will be the same for all the cubes produced .

The time interval  is the size of the cube boxes on the Z axis . In others terms, the time step that we are going to apply to our analysis. In our example , the time step of the forecasts is 3 hours . A size 12 hour box will produce 16  time intervals in Z. A box size of 24 hours, will produce 8.

One of the constraints of the cube is it must contain, at least 10 time lapses.  Therefore we will choose so a time lapse of 12 hours .

The distance interval   is the cell size in the XY plane.

If you do not select one or both intervals, the program will calculate one by default .

Visualization of the space-time cube

To visualize the cube you must Download a Toolkit containing two tools: one that allows to visualize the cube in 2D and another that allows to visualize it in 3D.
The Toolbox East downloadable at

Once downloaded and decompressed, you can add it in the geoprocessor by clicking on the ribbon Create -> Toolbox

We will see the visualization in 3D.

You must have a 3D window (Scene) in your project . If you do not, please make sure to, click  Insert -> New Scene

You must remove the reference to a layer of elevation. If you do not, the columns of the cube will be displayed relative to the ground level and you will loose visibility of the correspondence of the different time lapses.

In order to do this, go to panel Content, click on your scene -> properties -> Elevation surface, deploy Ground and delete all services present.

Execute the visualization command of the spatial-temporal cube in 3D:


Point on the cube that you just created. Submit a name to your viewing layer.

In the variables to display, you will not have, at this stage anything but  Count . In our next article we will discuss how to introduce other variables after having used the command analysis of emerging hotspots.

After some calculations and display process display , depending on your graphics card , you will have the result :

You can navigate in 3D on the stage to see where the time lapses with larger gusts (over the fixed threshold). But remember that it is only a visual Evaluation.

To determine statistically the number of overruns, the order   Analysis of  emerging hotspots  allows a  standard deviation ranking of the results.

We will discuss this topic in the next article.

Integrated Coastal Management: location (spaces) and localization (space)

Unlike in a previous article (Spatial Data Analysis and Spatial Analysis of Data) where nothing was specific to coastal management, in this article I will examine specifically Coastal Management. We will start by discussing how geomatics applied to the littoral zone displays very specific features.

Continue reading “Integrated Coastal Management: location (spaces) and localization (space)”

QField: the mobile device of QGis for Android

QField allows you to do fieldwork on QGis projects with Android phones or tablets.
In this article we will discuss how to install QField on a mobile device and how to install and work on a QGis project. Continue reading “QField: the mobile device of QGis for Android”

How to create a SpatiaLite database with QGis

If you start to feel tired of managing your data as shape files (this format dates back to the eighties!) and you yearn for a genuine database for your projects, most certainly, you will be looking to use PostGis. There is a big gap, and the ‘jump’ is, by no means, an easy one. In this article, we will discuss an alternative solution, the SpatiaLite database. And, believe me, it is much more easier…. Continue reading “How to create a SpatiaLite database with QGis”

How to rectify the geometry of a Postgis table

We have already discussed the topic of geometry validation through a series of articles (in french):

In this article we will discuss, in detail, how to detect and rectify geometric problems in a PostgreSQL / Postgis table using SQL queries.
Continue reading “How to rectify the geometry of a Postgis table”

Start with PostgreSQL / Postgis -Introduction to pgAdmin 4

Following up a previous article (Débuter avec Postgres/Postgis ), we will address an introduction to Postgres/postgis database management. We will load a shapefile, connect and load the Postgis layer from QGis.

The most  suitable method to manage PostgreSQL databases is by using the pgAdmin4 GUI.

This tool is setup automatically during PostgreSQL installation. You can launch it from the programme bar:

Continue reading “Start with PostgreSQL / Postgis -Introduction to pgAdmin 4”

Geoserver avanzado: el teselado (puesta en práctica)

En el artículo anterior (Geoserver avanzado: las teselas (principios)) vimos los diferentes conceptos para el teselado de mapas.

En este artículo veremos cómo implementar estas funciones en Geoserver. Continue reading “Geoserver avanzado: el teselado (puesta en práctica)”

Advanced Geoserver : tiling (quick start)

In the previous article (Advanced Geoserver: tiling (principles) we have explored the different concepts regarding the tiling mapping.
In this article we will discuss how to implement those applications using Geoserver.
In order to run the tiling, Geoserver uses an external module called GeoWebCache. Continue reading “Advanced Geoserver : tiling (quick start)”

Use of Landsat images (free) in your GIS

When we think of accessible data to integrate into our GIS, we hardly think of satellite data. And yet, it is possible, and just a few clicks away.

Let’s see an example. The following images represent the state of the vegetation for the islands of the Gulf of Morbihan, on the left for July 2014, on the right for March 2015. If we wish, we can obtain images every 15 days, the most just one week old.

ndvi juillet morbihan

ndvi mars morbihan

Findin, recovering and processing them in just 5 minutes. So, why deprive yourself?  Continue reading “Use of Landsat images (free) in your GIS”