If your organization has organized its information collection process, and determined the Essential Elements of Information (EEIs) necessary to make decisions, you can move on to developing the format for data collection.

It is not uncommon for organizations to collect information during a crisis using a completely different format from a neighboring jurisdiction or partner agency. The middle of a crisis is not the time to try and make a significant change to how you are collecting data. If working with a Digital Volunteer Network or Tech Sector partner, determine if they can provide applicable crowdsourced data in a format that aligns with your traditional sources. After the incident, it will be important to determine if a data standard can be leveraged to collect data on your EEIs.

Spending some time with your stakeholders in advance to determine the approach for data collection, data format, data attributes, and data sharing methods will save you considerable time converting both crowdsourced and traditional data sources to a format that your organization will be able to use. A significant amount of time is spent during each incident working to deduplicate data entries and crosswalking data fields between various sources. There are a few approaches that can assist with advanced planning to reduce this unnecessary process.

1. Does a data standard or model exist? If the EEI you are collecting information on already has a data standard that has been developed, attempt to record crowdsourced data using this standard format. An example would be the collection of community shelter data using social media. The National Shelter System is an existing standard to collect this information and is widely used by emergency managers at all levels of government. Even if all data fields are not used, ensure that the data fields that are collected align with the data standard. An effort to catalog existing data models can be found at FEMA Emergency Support Function Leadership Group's Model and Data Inventory. NAPSG Foundation has developed the Implementation Guidance on Information Sharing Standards for Crisis Management and Mutual Aid Technology as a as a simple guide on choosing the appropriate information standard for a given need, as well as, identify some key aspects for communicating that information between systems. Also take a look at the Tools & Resources page on this toolkit for links to other data standards & models.
2. Have common status attributes been identified? The next common problem in leveraging crowdsourced data during an incident is the use of attribute data that does not align with traditional data sources. An example would be the collection of power status of long term health care facilities. If emergency management is collecting power status information based on a status of no power, generator power, commercial power, crowdsourced data sources should also collect data in this same approach. Standardizing the status attributes will assist in the comparison of traditional and crowdsourced data products. One resource to assist in developing status attributes for your EEIs is the National Information Sharing Consortium's (NISC) EEI Publication Guidance. While the included recommendations provide a good first draft of attributes, it is important to develop the final attributes based on your agency's requirements in conjunction with any stakeholders that will be using the data. More information on NISC's approach can be found on their website.
3. Can the data be collected in an editable format? If it is not possible to leverage a data standard or common status attributes, the next best option is to share the data in a format that enables quick and automated conversion. One common pitfall for information sharing is the sharing of PDFs full of important data. A PDF does not allow for data to be converted into a format to compare traditional and crowdsourced data in the same tool or product. Work with a Digital Volunteer Network or technical staff at your organization to determine the best data format to share traditional and crowdsourced data among partners.
4. Is it possible to enable real-time data sharing? One benefit of crowdsourced data is the dynamic nature of real-time changes and adjustments from the crowd and applicable sensors. In some cases it may be possible to obtain a fairly accurate picture of current conditions during an incident. To leverage this functionality, crowdsourced data should be provided in a feed or service that can be consumed in real-time through an application programming interface (API). Manual uploading or emailing of files can add an unnecessary lag in time to capture EEIs and increases the chance of misinterpretation when comparing live reports from the field with traditional and crowdsourced data.