We’re all familiar with the concept of “point in time” or “snapshot” data. This kind of information has value, primarily because it’s easier to produce and easier to digest. But, this convenience comes at the expense of context.
Snapshot data by definition isn’t real-time data. But that’s not necessarily a problem. For most applications, real-time data isn’t necessary, and in a lot of cases it doesn’t even exist. No, the issue with snapshot data is that it necessarily leaves out context, by which we mean that snapshot data can’t tell you about trends, can’t help you identify patterns and can’t help you make relative assessments. In short, snapshots give you the trees without the forest.
We’ve separately discussed the art of data disambiguation and data normalization, but what is gained from all this effort? In short, it gives you the basis for data analytics. When data is well organized, it is easy for computers to compare large amounts of data and provide you with highly summarized results. Moreover, when a normalized and disambiguated dataset spans many years, you have the ability to look at how things change over time – patterns and trends – arguably the most powerful form of data analytics.
The Medmeme database, which, in addition to being normalized and disambiguated, now holds ten years worth of data, is an excellent example of the useful insights you can glean when datasets can be analyzed over time. Consider:
- Centers of excellence for specific diseases can shift over time, and with change over time analytics, you can test to see if a center of excellence is waning in importance and influence while also identifying where new centers of excellence may be emerging
- You can assess whether a KOL may be poised to lose influence based on published research and conference presentation trends over time or even spot a gradual move into new research areas that might indicate a shift in focus
- You can watch the ebbs and flows of research output and conference discussion over time for any disease – at a country level or globally – to see how research priorities and resources are shifting to gauge market opportunity
- You can watch for “hotspots,” research areas undergoing significant growth in activity; and again, you can’t look for growth without knowing prior levels of activity
- You can analyze share of scientific voice over time for any product or product class, perhaps comparing it to third-party data on such things as advertising spend to potentially identify key levers to maximize your own marketing and promotion planning
As you can see, the analytical possibilities are endless. Hopefully, this has got you thinking about the incredible insights that can be mined from change over time data, both to ascertain what happened (useful) in addition to what is going to happen (extremely useful).
Analyzing data in context – holistically – is the way to extract the most value from a dataset. But the quality of your analysis can never be better than the quality of your underlying data. That’s why to see the big picture, you not only need the sweat itself, you need a data provider that sweats the small details so that the outputs of your analyses are trustworthy and clear.