Enterprises are most successful when they treat data like a product. It enable to use data in multiple use cases. However data product should be designed differently compared to software product.
A Data Officer need to confront complexity of modern enterprise thru multiple data models. Using knobs data officer can enable management and governance of data. Knobs enable interpretation of data, build logical understanding. Most importantly knobs act as level using which data leaders can control how information is applied in various business process.
Use knobs to define how data flows across the applications. Limit flow of highly confidential data. Use knobs to anonymize and sanitize your data.
Using knobs publish non confidential data to internal and external audience. Depending on content of sharable data system generate right UX for consumers
Add more data points, diversity, variability to build dataset that represent world. Generate new dataset to handle cold start problem or test model
Check how compliance augmented and generative datasets are Check blind spots in NN behavior trained on these dataset and what can go wrong
From the blog
To build a commercial data product, create a base data product. Then add extension to these data product by adding various types of transformation. However it lead to complexity as you have to manage Data Lineage. Use knobs for lineage and extensibility
In complex problems you have to run hundreds of experiments. Plurality of method require in machine learning is extremely high. With Dataknobs approach, you can experiment thru knobs.
Data is new Intellectual property (IP). Data generated thru business process is important, Augmented data for Macine Learning, Data created , transformed to fuel new use cases in AI is also an IP.
Generality of AI product depends on variety of data you you. To build universal dataset for AI product you need many approaches to build gold dataset. Data knobs enable you manage and generate datasets.