Understanding a market or a phenomenon has to account for when something happened.
If you build an analysis that says 20% of records were published with missing data, the inevitable next question is “how do we put that into context”?
In other words, is 20% bad or good?
That will mean conducting the same analysis at a different time. It might be an hour earlier or a year, but the point will be to compare the data to a historic event in order to create some meaning.
If the number of records with missing data was 50% a month ago, then 20% is good. If it was 2% then 20% is really bad.
Your team, your directors, and the public will want this context when you publish data.
But that means you need the data, the data for today, and the data for a month ago. Or the data for a year ago.
But that’s not all.
If you’ve done a good job and found analysis that you think is useful, then you’re going to need the data in the future too. You’ll need the data in a month’s time or a year’s time.
To that end, good analysis is like a marriage. It is a long-term commitment.
In business terms, building analysis isn’t about completing a task, instead, analysts should think about building a service.
A service that provides a regularly updated source of data that is reliable and dependable. A service means a data pipeline combined with validation, tests, and a way to examine the data. From raw source to presentation, you need to think about what it will take to process the data on a regular basis.
Good analysis is a commitment that takes effort and skill.
Governments and businesses can use our platform and skills to build a reliable flow of data ready to be analysed in any way you like.
To find out more, get in touch.