Importance of Data Quality

The ongoing issue of homelessness does not seem to carry a large relation to the way data is handled, but there are multiple reasons why data can impact how homelessness in the U.S. can be solved. Shelters that we have worked with in this growing project all carry a dataset which contains sensitive and important information about each individual who enters the system.

There are a large number of fields which are used to learn more about each client and about the overall issue of homelessness itself. With the presence of so many fields there are multiple issues that spawn from various data. These issues impact crucial data that helps in creating a solution to homelessness. Some examples are given below, along with how their solutions can aid our cause.

One issue from the dataset we received involved information that was incorrectly inputted or gathered during the period of interrogation. Whether intentional or unintentional, this presence of incorrect data can cause those who research homelessness to study false information and shelters to handle inaccurate data.

Conflicting records involving reported shelter enrollment without the presence of an exit record can cause repeating sets of data for one client. These repeating sets are spacious and unnecessary, and they impact the overall quality of data on each individual client.

Families that have visited the shelter in the past visit the same shelter once more, but a few family members get mis-recognized as new clients and some members are given their only personal ID back whereas mis-recognized members receive new personal IDs. This not only causes confusion amongst the clients, but also creates repeating data sets which are not needed.

Many shelters do not share their information with each other, causing many sets of information to be scattered and repeated. What is needed is one set of accurate information that can be shared with all shelters to ensure the most effective flow of information and use of space.

There are instances where clients enter inaccurate information into multiple data fields. This type error needs to be corrected because it will prevent inconsistent datasets, and aid in improving the data belonging to many different homeless shelters.

Written by Michael Losritto