9 tips to ensure high-quality ArcGIS Open Data downloads
It’s Open Data Day on March 5th. To celebrate, we’re running an Open Data blog series this week and next, featuring posts that explore the concept of freely publishing public data and how it relates to GIS technology. In the first post of the series, senior support consultant Jonathan Nowlan imparts nine tips to ensuring high-quality open data downloads from ArcGIS Open Data.
Sharing data with open data is easy. However, it’s important to spend time preparing your data to make sure that you’re offering high-quality data downloads to your users. Below are some best practices to ensure your data is open data ready.
When it comes to sharing feature services, one of the very first considerations you’ll want to take is to decide how your services will be organized. Does it make sense to publish each layer individually, or can certain layers be bundled into a multi-layer service? While each layer within a multi-layer service will appear as a unique layer in ArcGIS Open Data, it’s important to understand how things like titles, descriptions and tags are derived with multi-layer services (described below). Also, there are performance considerations when sharing multi-layer services. To keep performance optimal, keep multi-layer services under 20 layers.
Multi-layer service as it appears in ArcGIS Online (left) and ArcGIS Open Data (right). Notice that the three layers are offered as individual downloads.
Max Record Count Under 5,000
When publishing a feature service that contains a large number of features it can be tempting to increase the maximum record count from the default value of 1,000. This parameter controls the maximum number of features that will be displayed in your current extent. If you do need to increase the max record count, we recommend using a value under 5,000. While you may wish to display your entire dataset at all scales, it’s not necessary with ArcGIS Open Data. All download requests will receive a full copy of the dataset regardless of the max record count. Keeping the max record below 5,000 will ensure optimal performance, not only for ArcGIS Open Data but also for ArcGIS Online.
Screenshots showing the maximum record count value as it appears on service URL (above) and the Service Editor window in ArcMap (below).
Add a Title
Sounds pretty basic, right? Titles are important. It’s the first thing a user will see when exploring available datasets on your open data site. Choose a title that’s informative and easy for everyone to understand.
For individual feature services, the title will be derived from the ArcGIS Online item page.
For feature services containing multiple layers, the layer titles will be derived from the layer names at the time of publishing. Title changes can be done locally (via ArcMap or ArcGIS Pro) by changing the layer name. However, you’ll need to republish your feature service for the changes to take effect on ArcGIS Online and ArcGIS Open Data.
In this multi-layer service example, the titles shown in ArcGIS Open Data would be derived from the names of the individual layers included in the service (Parks, Trees and Water bodies) and not the Feature Service item title (My Areas of Interest).
Don’t assume your user will be satisfied with a good title. Give them a clear description of the data avoiding the use of internal \industry specific jargon, or terminology that may not be clear to all users.
For individual feature services, the description will be derived from the description provided on the ArcGIS Online item page.
For feature services containing multiple layers, the layer descriptions shown in ArcGIS Open Data will be derived from the individual layer descriptions provided at the time of publishing. Description changes can be done locally (via ArcMap or ArcGIS Pro); however, you’ll need to republish your feature service for the changes to take effect on ArcGIS Online and ArcGIS Open Data.
For multi-layer serivces each layer within the service will inherit the description provided for the layer in ArcMap or ArcGIS Pro.
Fields & Field Names
Have you ever received a dataset that contained field names that looked like some sort of alien language? Don’t be the purveyor of confusing data. Your fields contain valuable information so make sure your users can quickly understand the data they downloaded. Here are a few approaches you may want to consider:
Change field names
While this may not be possible for every organization, you can use the Alter Fields tool in ArcGIS for Desktop or ArcGIS Pro to change the names of fields that may be confusing.
If changing the field name is not possible, you can specify field aliases to show a more descriptive and informative field name. These aliases will display in ArcGIS Online and ArcGIS Open Data, but they will not be included in downloaded copies of the data. Downloaded copies will contain the original field name.
Hide unimportant fields
Sometimes less is more. Don’t overload your user with useless information. Before publishing feature services consider hiding fields that contain information that’s only valuable to your organization. Give users meaningful fields with information they’ll be able to use.
Regardless of the approach you take, we recommend including metadata about your fields so users can quickly find out information about the field if needed.
In this example, we're looking at the layer properties in ArcMap and we've added an alias to the field "eg_decid". "Evergreen or decidious" will be easier for users to understand. We've also decided to turn off the field "Date last trimmed" since that information was deemed to be internal.
Tags often seem trivial, but try to spend a few minutes thinking of tags (keywords) that can be used to help users search for the dataset. This will ensure that the dataset can be easily found during searches. Tags can be added on the ArcGIS Online item page for the dataset.
Tell your users what they can and cannot do with their downloaded data. We recommended that you choose a Creative Commons License. Licenses can be specified on the ArcGIS Online item description page. If you don’t include a license you’ll receive a warning regarding that dataset when looking at the Data Reports page
Each dataset made available to your Open Data site will contain a thumbnail. By default this thumbnail will be derived from the ArcGIS Online item page.
However, you can modify these thumbnails and replace them with thumbnails that are more appropriate for your organization or that suits your web design aesthetic. This blog post is a great resource for those of you who may be looking to customize your thumbnails.
I’m going to go out on a limb here and say that metadata is not the most glamourous or exciting aspect of GIS data. However, it plays an important role in ensuring the quality of your data. Whether or not your organization requires you to adhere to a specific metadata standard, it’s worth spending the time to include metadata with your data. You can find more information on how to work with metadata in ArcGIS Online here.
Following these simple guidelines will ensure you have a professional, accessible open data site with high- quality data that will hopefully empower your users.
Happy Open Data Day.