I have been presented with the problem to predict the total number of employees that will reside in each industry in Colorado in 2030 based on the current data that is available. All of my data sets are available for download from my dropbox online so if you check the .ipynb files there should be links to my data or csv files. I have created a few different models to test and see what works with the best some with normalized data and some with it not normalized. I would like to give credit to samarthkadaba for the SVC_Neural_Network file. I used a substantial amount of their work to fit my model. I did create my own in the end to fit my data better but I wanted to say that here.
Government agencies and nonprofits, particularly the US Bureau of Labor Statistics and groups with an interest in education are often attempting to forecast what jobs or careers, and what specific skills will be in demand from employers ten years into the future. This is an example offorecasting or extrapolation. One must make a prediction about what will be the changes indemand for jobs and also what will change in those specific jobs. For one thing the predictions by the US Bureau of Labor Statistics are that jobs for mathematicians and statisticians will be increasing at a rate much faster than average in the next decade. Secondly, these predictions and the under lying analysis are a great example of the kind of work that mathematicians and statisticians do.
- A forecast of what jobs / careers will be in demand for Colorado in 2030 (and how that demand will change from 2020).
- A prediction of what skills Colorado employers will be needing from employees in 2030 (and how that will have changed from 2020).
- An analysis of whether the educational institutions in the county are currently ready to meet that demand, or what transformations are needed. This problem is one that is done frequently at the national level and you should be able to findsome documentation of that. It can be done at the level of a single county however data may beharder to find, and there is likely to be much less certainty about predictions.