Data powers predictive and prescriptive analytics. As enterprises rush to mine their data, AI and ML adoption is on the upswing. A recent survey by Dataconomy suggests that 75% of C-suite executives have AI and ML initiatives on their agenda. However, over a third of these projects are bound to fail, as companies struggle with poor data quality. Inaccurate and duplicate data leads to miscalculating demand and targeting the wrong prospects. This is confirmed by 59% and 26% respondents respectively. Unifying unstructured third-party data, semi-structured data or data from relational databases remains a challenge. Data silos also reduce accuracy of data driven decisions.
Almost half the data scientists report spending over 10 hours per week on data preparation. The other half spends upwards of 40-hours weekly on data preparation. Given the scarcity of data scientists and the value of their time, this is surely another big drain on a company’s resources.
The challenges for AI and ML as well as analytics are
- data exists in different systems (28%)
- requires merging from different sources (27%)
- needs reformatting (25%).
Do let us know if we can help improve your data models and improve the productivity of your data scientists with better data.Tweet
Named Entity Recognition and Entity Resolution »
« Merry XMas and a Happy New Year!