Computer Science Application
Primitive Datatypes Solutions
B.6 Project Deliverables
The project will have one deliverable, the finished analysis application.
B.7 Implementation Plan
The implementation plan focuses primarily on the two phases and the rollout of the prototype
and final application. At the end of each phase, the application and all associated resources will
be rolled out to customers. These resources comprise of a brief training video series and two
weeks of initial 24/7 support. We will also include a set of training data to ensure that no use
trains on a production dataset.
In terms of access to the software, we will restrict the code repository to only members of the
assigned development team. This will prevent any hallway testing or edits to the code base by
other employees of analysis. This will help ensure proper change control management and ensure
that no developer can push changes without approval. If needed, we will provide a tool to
convert .DB files or .SQL files to CSV files so that the data can be properly changed. This
conversion tool must be able to also convert data back to the original file format and be easy to
use.
B.8 Evaluation Plan
The evaluation plan will focus on how well the program is implemented into existing daily
actives. We must provide surveys and collect data from super users on statistics such as how
many times a user opens the application daily. To ensure that our tool is providing useful data
protection, we must also track use cases and ensure that proper data is being inserted into the
analysis. As mentioned before, the program should provide a meaningful service without
interfering with sensitive data that could interfere with HIPPA or other data compliance.
Additionally, we need to confirm that the noise being added to datasets does not go beyond the
existing data or create errors. To accomplish these tasks, we will need to initially track usage,
data, and noise statistics for customers that use the product during the first deployment phase.
This data must be stored locally with the client and transferred in a secure way. Once we are
certain the application is working as intended, this data should be destroyed, and stakeholder
sign off should begin.
The initial goal of our testing will be to determine as many errors and data discrepancies as
possible. As the amount of errors fall into an acceptable range and deployment is solidified, our
testing focus will then switch to ensure a level of data privacy is maintained. To accomplish this,
we will ask for periodic customer reports on data statistics that does not reveal any sensitive
information to us. Our team will analyze these statistics for future updates and rollouts
