Data Questions to Consider - Office of the Provost

Explore More

Before doing an IT implementation or, even better, before going out for proposals, the Data Governance and Support Office would like to suggest the following questions to consider. Think of these as the “data requirements” to complement the “functional requirements” and “technical requirements” that are usually gathered as part of a project. It will be very helpful to be able to assess whether an application will meet your data needs aside from the business requirements.

Questions? Suggestions? Let us know.

	Title	Question	Category
1	Data capture	Does a central repository exist for this type of data? Or are the data captured in spreadsheets or shadow systems?	Data collection
2	Data entry	How do data need to be entered/updated? Bulk upload? Integration? User entry?	Data collection
3	Duplicate values	Is there a potential for duplicate data entries? How do you reconcile duplicates?	Data integrity
4	Data Cleanup	Does the data have to be evaluated or cleansed for accuracy? How does that need to be done?	Data integrity
5	Data validation	Is there a data validation process? Will a data validation process be necessary?	Data Integrity
6	Data Retention	Does data ever get archived or removed from the system? How long must you keep the data?	Data Integrity
7	Unique IDs	What unique identifiers exist for these data? How is this unique identifier generated? (Is it a sequential reference ID or combination of data attributes? Is there a naming convention or logic associated?)	Data Integrity
8	Unique ID reuse	Are unique identifiers ever reused? Why?	Data Integrity
9	Error handling	Do you have/will you require a process for handling data errors? If so, describe the process?	Data Integrity
10	Updating the data	Do you have / will you require a data update control process?	Data integrity
11	Data migration	If adopting a new system, what data will migrate into the new system? Historical Data Current Data New Data All of the above	Data Migration
12	Data permissions	Do you have / will you require a data sharing control process? What workflow is necessary to support that process?	Access Management
13	Data Structure	How does the data need to be structured and presented for analysis (longitudinal, cross sectional, fiscal snapshot)?	Analytics
14	Frequency/ Timing	How readily available does the data need to be to meet the needs of the business (e.g. real-time, end of period, once a year)?	Analytics
15	Basis of observation	What is the basis of observation (person, object, task)? What is the focal point of the data?	Data structure
16	Missing Values	How do you treat missing values?	Data structure
17	Metadata requirements	What information about the data do you have or want? Descriptive (such as keywords, helps people find things), structural (such as relationships, helps people understand how things are organized) or administrative (such as update dates, helps people understand how things are processed)?	Data structure
18	Data harmonization	Is there data manipulation required to make the data compatible across systems? Where can/should that be done?	System Requirements
19	System Restrictions	Are there any system restriction for downstream systems that we will have to accommodate?	System Requirements
20	Definitions	Do you have documented definitions?	Data quality
21	Quality dimensions	What components of data quality are most important to you? Accuracy Completeness Timeliness Consistency Availability Uniqueness	Data Quality
22	Provenance	Are these data the authoritative source for the information? For all elements or a subset? If not, what is the authoritative source?	Data Quality
23	Data Familiarity	What is the level of data familiarity of our end users? Do they use spreadsheets? Are they comfortable with having access to the database and creating queries? Do they have experience with ad hoc reporting and selecting criteria to create custom reports? Does the user just want to push a button that extracts the data in the necessary format?	Use