5.5.3. Analysis 1 - Accessibilité

L'analyse d'accessibilité permet de calculer la distribution spatiale du temps de trajet depuis ou vers le ou les établissements de santé les plus proches pris en compte dans l'analyse, en tenant compte si besoin d'un temps de trajet maximal spécifié.

Cette analyse prend également en compte les contraintes et barrières aux mouvements, ainsi que le tableau de scénario de voyage qui définit les vitesses de déplacement pour chacune des classes d'occupation du sol fusionnée. La distribution spatiale de la population cible n'est cependant pas prise en compte dans cette analyse car elle n'a pas d'influence sur le temps de trajet.

Pour en savoir plus sur cette analyse, vous allez créer un raster de distribution du temps de trajet de deux heures pour 10 établissements de santé situés dans une zone du Malawi, en utilisant la couche d'occupation du sol fusionnée créée au chapitre précédent.

Pour cela, sélectionnez l’analyse "Accessibilité" dans le module "Analyses", puis remplissez les différents champs comme indiqué ici:

Données en entrée:

(1) "Sélectionnez la couche d'occupation du sol (raster)": sélectionnez ici la couche d'occupation du sol fusionnée ("couverture terrestre fusionnée [couverture terrestre fusionnée]") que vous avez produit comme indiqué dans la section précédente ci-dessus.

(2) "Sélectionnez la table de scénario (table)": utilisez le menu déroulant pour sélectionner une table de scénarios de voyages existante importée dans le projet. Une liste déroulante vide indique qu'aucune table de ce type n'a été importée (exemple ci-dessus).

(3) "Scénario de voyage à utiliser": utilisez cette section pour définir la vitesse de déplacement et le mode de transport (à pied, à vélo, à moteur) rattachés à chaque classe de la couche de d'occupation du sol fusionnée en:

• Important le contenu de la table de scénario sélectionnée si une telle table a été importée dans le projet (voir le point (2)), en cliquant sur l’option "Importer le contenu de la table de scénarios sélectionnée ".
• Éditant manuellement les informations reportées dans les colonnes "label", "speed" et / ou "mode" en double-cliquant simplement sur la cellule correspondante. C'est un moyen pratique de créer rapidement ou de modifier la vitesse et/ou le mode de transport sans avoir à modifier une table Excel externe. Veuillez noter que les informations finales entrées dans cette table au moment de l'exécution de l'analyse seront enregistrées dans une nouvelle table de scénarios avec la classe "scénario traité" et les mêmes tags que ceux saisies au point (9).

(4) "Sélectionnez la couche existante des structures (vecteur)": utilisez ce champ pour sélectionner la couche contenant les localisations des structures de santé à utiliser dans l'analyse.

(5) "Structures sélectionnées": le contenu de la table attributaire de la couche des structures de santé sélectionnées apparaît dans la section du panneau. Cette table vous permet de sélectionner les structures de santé que vous souhaitez prendre en compte dans l'analyse. Toutes les structures sont sélectionnées par défaut, mais cela peut être modifié en décochant les installations à l'aide des boutons de vérification de la colonne "amSelect" ou en utilisant des règles de filtrage/sélection utilisant un ou plusieurs filtres à l'aide de l'outil de filtrage situé en haut de la table. Cet outil de filtrage est simple à utiliser: on peut construire une règle de filtrage en combinant n'importe quelle colonne d'attribut avec des opérateurs et des valeurs associées (par exemple, "capacité >= 1000", "hf_type == Centre de santé").

The columns presenting a label starting with “am” are not coming from the attribute table, but have been generated by AccessMod. These particular columns are the following:

- "amSelect": allows you to select/unselect the facilities to be run in the analysis

- "amOnBarrier": informs whether the health facility is located (“yes”) or not (“no”) on a barrier. The analysis cannot be performed if health facilities are located on barriers. If one or several of the health facility(ies) is/are located on a barrier, you can do one of two options:

• Decide to proceed without the facility(ies) in question by un-selecting them from the "amSelect" column and proceed. However, beware that in this case the results from the analysis will not consider the entire health facility network.
• Correct the health facility layer to move those located on barriers (see Section 3.3.1.6). This operation must be completed outside AccessMod, in your GIS software and the new layer must be imported in AccessMod.

- "amOnZero": informs whether the health facility is located (“yes”) or not (“no”) on a landcover category that has a speed of zero set in the travel scenario table. In this case the analysis cannot be performed, and you can do one of two options:

• Decide to proceed without the facility(ies) in question by un-selecting them from the "amSelect" column and proceed. However, beware that in this case the results from the analysis will not consider the entire health facility network.
• Correct the corresponding speed(s) of travel in the travel scenario table, by giving them a value greater than zero.

- "amCatLandCover": provides information the land cover class, i.e. the pixel value in the land cover raster layer. 

(6) “Type of analysis”: Select the type of analysis you want to conduct: isotropic (ignores the DEM and is slope-independent) or anisotropic (uses the DEM and is slope-dependent). In this exercise, we will select “Anisotropic” under "Type of analysis". This implies that the DEM is used to compute slopes, that are in turn used to modify the speed of travel indicated in the travel scenario table when either the "walking" or "bicycling" travel modes have been chosen for at least one landcover category. Refer to Section 3.3.2 for details on how the slope affects speeds of travel. If the "isotropic" type of analysis is chosen, slopes have no effects and the speeds of travel are not corrected.

(7) “Direction of travel”: Choose the direction of movement considered for the patients, either "From facilities" or "Towards facilities". This choice is only visible when an anisotropic mode has been chosen, because the direction of travel can influence the time of travel. To do the exercise, choose "Towards facilities".

(8) “Maximum travel time (minutes)": Specify the maximum travel time (in minutes) after which the analysis stops. For the exercise, specify 120 minutes. Specify “0” in this field if you want to compute travel time for the full extent of the study area, i.e. with no limit of travel time.

Important note on maximum travel times

Starting with version 5.2.4 of AccessMod, travel times are only given in integer numbers of minutes in all output with travel times (raster and tables). This was done to minimize the size of travel time raster files. If you specify "0" for the "maximum travel time", the maximum travel times that can be computed is actually 32'767 minutes (i.e. 22 days, 18 hours and 7 minutes), which should cover the countries in the vast majority of cases. If output travel times exceed this limit, the corresponding cell values in the output travel time raster will be encoded with 
"-1" in order to be easily identified when the raster is displayed in a GIS. However, if you need to compute travel times beyond the limit of 32'767 minutes, you can specify a higher maximum travel time (e.g. 50'000), which will change the format of the output travel time raster, and the maximum upper limit will become 2'147'483'647 (about 4085 years). But note that in this latter case the size of the output raster will increase.

(9) “Add short tags”: Give short tags to be attached to the different outputs of the analysis. We will use "accessibility 120" for the present exercise.

Validation:

(10)  The validation section tells you if everything is fine for you to start the analysis. First it gives you an estimation of the required memory and disk space necessary to run your analysis. Note that these estimations are indicative, and the real requirements may vary. If the memory requirements are close to the memory allocation you gave to the Virtual machine, it is a good idea to increase the latter (see Section 4.2 for how to do that). If anything is wrong in the settings, the "Compute button" will be red and a warning/error message will appear in the list, indicating errors that you will need to address prior to running the analysis. If everything is fine and all the fields have been correctly filled, a green “OK” will appear next to each item in the list and the "Compute button" will be grey (example here above). When this is the case, you can hit the button to launch the analysis.

 A transparent window with some text and a progress bar will appear in front of the panel while the analysis is being conducted. Please wait until this window disappears to continue using AccessMod. Once this is the case, go back to the Data module to check the four output data sets that have been generated:

A filter is automatically applied in the Data module when you return after using one of the tools. The application of this filter results in having only the latest output datasets appearing in the list of available data. This simplifies the check of these data as well as their selection for exporting them outside AccessMod.

If you want to see the complete list of data currently stored in the virtual machine, just remove the tags appearing in the filter section on the left-hand side.

The accessibility analysis generates the following two datasets:

1. scenario processed class: Table containing the travel scenario that has been processed.
2. travel time class: Raster format layer containing the spatial distribution of the travel time expressed in minutes.

Before being able to open and look at these datasets you need to archive and export them outside AccessMod (i.e. outside the AccessMod virtual machine). To do that, do the following steps in the "Data" module:

-       Check the four datasets in the "Available data" section of the panel.

-       Go to the "Archive" section, enter a prefix name for the archive, and hit the "Create archive" button.

-       Once the archive is created, select it from the dropdown list below (your most recent archive will be at the top of the list), and hit the "Export archive" button.

-       The archive (in zip format) is then automatically downloaded in the default Download folder used by your browser. Your browser may ask you to specify the location where to save the archive.

-       Go to that folder and unzip the archive to access its content.

 We are now going to open these different output datasets. Start by opening the Excel file that is inside the "table_scenario_processed_accessibility_120m" folder that has been unzipped from your archived ZIP file.

 This is the table containing the travel scenario that has been applied during the analysis:

If needed, it is possible to modify this table in Excel and re-import it in AccessMod as a new scenario table (we are not going to do that now).

The exported raster layer (travel time) is stored in ERDAS Imagine format in the corresponding folders that you have just unzipped. To visualize it you need to use a GIS software (QGIS, ArcGIS, GRASS).

 Importing the .img file containing the travel times to the nearest facility (stored in the "raster_travel_time_accessibility_120m" folder) would result in the following layer in QGIS:

This raster format layer is the result of the accessibility analysis and shows zones in which the population can reach the nearest health facility in two hours or less. Pixel values are the travel times in minutes and the colors in the Figures are the ones by default in QGIS to visualize them. Colors can vary depending on your default color scheme. Note that this screenshot also contains the location of the health facilities used in the context of this analysis (red dots).