“Data mashing” seems like a poor description of a process. You are not bludgeoning the data. Instead, data mashing is more like using a kaleidoscope – mixing together different types of data and seeing what it looks like through a single lens. The difference is that, instead of using mirrors to create patterns, we rely on visualizations to expose them in our data. Activities like juxtaposing, pivoting and charting can provide us with new and valuable insights.
Beginning in the 1990’s the emergence of MP3 files allowed mixing different musical tracks; and from there, digitized mashups began to extend beyond music to video, web services and even data analytics. The growth of big data, cloud services, and free and open source software platforms like Python and R have made it easy for individuals to create data mashups and, in effect become their own data scientists. The website ProgrammableWeb, for example, has a directory of over 6,000 data mashups.
As data mashups have become more commonplace, they have evolved into a resource for businesses looking to monetize their existing data assets as a data product. But some other benefits that data mashups can offer organizations include:
- uncovering new revenue streams
- developing efficiencies and best practices
- revealing hidden relationships
- improving decision making
The bad news is that the mashup landscape is still somewhat free range. Data quality and availability can be challenging. The good news, for enterprises at least, is that they already have deep investments in their own proprietary data.
This could comprise transactional data, like sales records, purchases, inventory, phone records, CRM and chatbot data. Data mashing provides a new way to obtain value from this “dark data” that has been accumulated but rarely used. Since much of it is already clean and structured, with an established data lineage, data scientists don’t have to do much preprocessing of it.
The term “data mashing” suggests an exploratory, even messy process, perhaps a first iteration that necessitates significant manipulation and massage of the data. But the latest generation of analytics tools make self-service mashups easy enough that even non-technical end users can obtain insights from their data. With Orbit, you don’t necessarily need a data scientist to do the mashing.
Given access to the right data sources (for example, in a well-managed data lake), end users can explore and discover insights themselves. Orbit even allows virtualized connections between data sources, so there’s no need to connect to a data warehouse or move data around to report on it.
You can combine application data from cloud and on-premise databases with unstructured data sources like Hadoop or flat files. Data from multiple, disparate sources in varying locations and formats can be used, without replication, to create a single virtual data layer for reporting and data discovery.
Orbit can help you obtain data insights across your entire organization. Learn more about data mashing in our whitepaper Transforming Your Data for Analytics: Three Options. Download your copy.