Preliminary Requirements


Identify Data Requirements

  • Identify the entities that will be required in the system
  • Start the process of data preparation
  • High Level Scope
  • Plan for data preparation
    1. Carry out a workshop to identify entities.
    2. Carry out data analysis to understand the quality of the data.
    3. Prepare a plan for cleaning up the data.
  • List of entities
  • Data quality assessment
  • Data Preparation Plan

During the Functional Decomposition activity, it is likely that a number of entities will be identified, These should be carried forward into this activity.

Before commencing the team should have a basic understanding of entities, attributes and relationships. Whilst the discussion is focused on entities, there will likely be some discussion around attributes.

The question that forms the core of data requirements is:

"What are the people, places and things we want to keep track of?"

Entities should be captured in the Excel Data Modeling Worksheet. There is no need to capture all the attributes or relationships at this point. Keep the focus on the Entities and any attributes that may be unusual or unique to your business.

The second part of this activity is to understand the quality of the data in your current database. This can be accomplished by working with your Database Administrator to investigate the quality of data in particular tables. It will also involve running various consistency checks to identify data issues. Also talk with staff who use the existing system on a daily basis. They are probably aware of data which they cannot trust.

It can be assumed that most of this data will be transferred into a new system. A decision needs to be taken by the Sponsor and Steering Committee what their goals are in terms of data quality. The decision will be that data should be somewhere between two extremes:

  • No better than what we have now with the exception that we will clean up whatever data will stop the new system from working
  • 100% accurate, consistent and complete

This decision is important as it will determine the data cleanup effort required. The decision may be on an item by item basis. We will clean up all addresses but not phone numbers. From this decision, a plan can be put in place for cleaning data and the data cleanup run as a separate project. The conclusion will need to be before the trial conversion in the Configuration and Customisation phase. Remember that no matter how thorough this analysis, there will still be data found that needs to be cleaned up. Time should be allowed for undiscovered errors.

This is only part of the work required to prepare data. In addition to cleaning up existing data, and filling gaps. you will probably need to:

  • Change the format of some data. For example split phone numbers into area code and phone number
  • Add data not currently contained in the existing system.
  • Reorganise data. This may mean grouping data by state or customer.

Most of this work will not be clear until a package is selected. If however all the work is left until the end, you will likely find the people required to fix the data are also required for configuration, testing and training. It is best to get obvious data cleanup done as early as possible.

Data Modeling Worksheet An Excel worksheet to capture information about data
Data Quality Assessment Record any data issues and track progress towards rectification
Data Preparation Plan Plan for the correction of data
Data Modeling User Guide Instructions on how to carry out a data Modeling workshop and how to record information.

Return to the top