Exposing Dirty GIS Data in ArcGIS Online
ArcGIS Online is a great product, but it can be limited by the quality of your data. After assisting several clients with launching their ArcGIS Online Organization, I have noticed a recurring theme. ArcGIS Online exposes “dirty” data. What I mean by “dirty” is the organization of their geodatabase: feature class names, aliases, attributes, data entries, and lack of domains.
How to Prepare your GIS Data
So before publishing your services, consider following these steps to prepare your GIS data for ArcGIS Online:
- Geodatabase Schema: A good way to organize your geodatabase is by creating datasets that create logical groupings of data and can be published as services. A good example of this is creating a Water Distribution dataset that has all feature classes associated with water infrastructure. Make sure the names of the feature classes are meaningful. Remember, while the names might make sense to you, other members of your organization, and even the public, are likely not going to know that feature class name “CSV” stands for Curb Stop Valve. This can be done by renaming the feature class or changing its alias. Check out the Local Government Information Model for a good starting point.
- Field Data Types: Once you have your feature class names then you’ll need to work on theattribute/field data types. For instance a field that is used to enter a date should be a field type “Date” and a field with only numbers should be a field type “Integer”, “Float, or “Double”. One of the great things about ArcGIS Online is the ability to create filters and out-of-the-box web apps. However, this can be limited by fields not being of the right type. Here’s an example: let’s say you had a field that showed cost and wanted to do a summary totaling the cost. This can only be possible if that field type is an “Integer”, “Float, or “Double”. If that field type is “Text” then entries in that field are not recognized as numbers and can’t be totaled. This step could potentially be very time consuming, but I promise that your efforts will be worthwhile.
- Field Names: The next clean-up step focuses on field names. The easiest way to accomplish this is to alsochange their aliases. The reason for changing the aliases of the feature classes and attribute fields is to simplify the web map creation process in ArcGIS Online. When a service is brought into a web map, it automatically names all the data layers based on their aliases. Which means, by cleaning up the feature class and attribute aliases, you shouldn’t have to spend time renaming the data layers in your web map or configuring pop-up field names. This could potentially save you and members of your organization a significant amount of time.
- Data Cleanup: Now that your fields use the proper data types and aliases, it’s also important to standardize the data in the fields. Correcting any typos and making sure all entries have the same syntax, for example: TN instead of Tennessee or vice versa. You might also want to consider creating domains, which enable drop downs for entering data in a field. This is especially important if you plan on doing any data collection/editing with ArcGIS Online via (Web App Templates, Collector for ArcGIS App, etc.). Having domains removes the possibility of someone who is performing the edits to make typos, standardizes data entries and expedites the collection/editing process.
- Feature Attachments: The final piece of this puzzle is to enable feature attachments, which allows you to attach pictures or documents to your data. If you’re using the hybrid approach to ArcGIS Online, then you will need to enable feature attachments in the geodatabase. If you are going to host your data using Esri’s cloud environment as “Hosted Services” then you can enable feature attachments after publishing the data as a service.
Worth the Work
In the short term, following the steps outlined above might create more work. However, the long term benefits greatly out way the short-term work. Web maps and apps will be easier to create. Data collection/editing will be standardized. Users across your organization and the public will be able to better understand your data.