Data Analytics in Government-Breaking Down the Barriers
The challenge for all organizations is change, but this is particularly true in government. This is in large part due to the need for government to survive and continue to function from administration to administration and through party changes. This same strength can be its greatest weakness as long term employees are resistant to new ideas, as they have seen so many come and go. This leads to the question of how can we shake up government and change the culture and outlook. The new world of data analytics is an opportunity to do just that.
With the advent of predictive analytics, we can put data in the hands of a caseworker, a parole officer, an educator, or any type of government employee who makes decisions that will alter the way they think on a daily basis. Let’s take the example of a child caseworker who has to make tough choices every day. Currently these caseworkers on an average stay in their job for three years for various reasons. This short average career lifespan means the built up knowledge that the long term case worker has, is lost on a regular basis. The real power of data is in bringing the knowledge of the twenty year caseworker to the brand new one in the field. The representation of this is in the difficult daily choices a caseworker makes about whether to remove a child from a home or not.
“If a case worker had the power of analytics on a tablet, the same data points could be used which would help reduce risk factors”
The obvious physical abuse cases are the ones where there is 100 percent certainty that the child should be removed from the home. Few fit into this category, most are in a grey area and nowhere near black and white. Imagine if the case worker had the power of analytics at their fingertips on a tablet providing them with not only percentage chances based on historical data, but with correlating factors that contribute to the risk. These same data points can be used to allow an application to connect the caseworker with the other state employees who can help reduce these risk factors. Some direct examples of this would be: if one of the parents is on parole their parole officer can be notified automatically, a child’s school counselor could also be made aware to ensure they are doing regular follow ups with the child, or if they are eligible for benefits, such as food stamps that they are not receiving which could help the child. The extra information could be collected onsite and the application can be made. These are just a few examples of how these correlating factors can be used to help improve the chances of a better outcome for the child.
So how do we get not only the agency responsible for child welfare, but the other agencies having pertinent data for the case worker to share it? There are several key components to doing this. Let’s walk through the big ones.
First and foremost you need to choose a problem to target, that everyone wants to solve. In this particular case who wants to say no to help improve children’s lives. Second, you need to be engaged in the entity. The particular problem is pertinent to from both ends of the spectrum.
You need to work with the top level executive management so they understand the power of the knowledge it can bring to not only them, but their workers. Involvement with those in the field who are truly getting the work done is also critical. You will need to work with these staff to understand not only what data you need, but what it is telling you. Without this intimate business knowledge all the data scientists with petabytes of information are useless.
Third, get a quick win. Start small by taking data sources from the agency. This use case is pertinent to publicly available data such as sources like, the census bureau or university studies. Although a limited subset of data may not give you the bigger win you are looking for it can give you some insights into specific issues in processes or practice. It is also useful, to identify data quality issues that you will need to overcome. And trust me there will be data quality issues. One of the great things about data analytics is that the same systems you use for the predictive analytics can also be used to perform data comparisons between data sets to improve data quality.
With these three key pieces you can get the basis needed to make an effective sales pitch to the other agencies, divisions, or departments whose data you want to improve your algorithms and create the big win that allows for the buy in on a much larger scale. Once you can get the first domino to fall, the others will follow once they see the power of predictive analytics.