Please consider taking a brief look at part of the project*
*Note that all visualizations are FAR from final products
"Tech" is very broadly defined here as anywhere from a componenet manufacturer, or software development company*
-
Data collected from WorldBank, consisting of over 45 countries, with 75 features, across 25 years
-
Language utilized: Python
-
Semi-large dataset
-
Unsupervised in that features and countries are labeled, but there is No Given Target Feature
- Display tangibility of Python utilization
- Focus is that of a consulting use-case, wherein Market Entry/Attractiveness will be the underlying goal
- Utilize the versatility of Python to reach insights that will contibute to the consultant's decision making and client direction.
- Heavily clean and organize the data in several ways
- Apply Machine Learning (Unsupervised) where possible
- Implement redundancy to display alternative approaches whenever convenient
- Avoid Pipelines (as irritating as that may be)
this is intentinoally ineffecient and visual for those less familiar with Python and Data Science
- Use consulting's MECE ideology in approach to solution
- Create MECE buckets wherein countries are ranked via their "tech market attractiveness" and classified accordingly
- Classification would be ideal, if we had a target variable
- Ensure that we utilize all of the data available
- In this case, Supervised Learning is substantially simpler (and quicker) than Unsupervised Learning
The more challenging a personal project is, the more impressive the result
The broadness of the goal and purpose, makes the actual process of this project a bit more tedious, however I believe it will pay dividends from an output standpoint.
Objective ambiguity resolution is more difficult to display when a project is being completed purely out of one's own interest. This is indeed an underlying purpose of the project.
Thank you for reading
Alexander