Problem Statement
To determine and forecast job growth and workforce in various industries in the United
Kingdom. This will help the United Kingdom Government and the Education industry in the
following ways:

• It will help in determining the training and course programs in the various industry sector based on job growth
• The education sector in the UK is a source of investment with investments of £33 billion annually through tuitions fees and grants and the government spends £87 billion. The forecast for labor and workforce in various industries will help the government to improve the allocation of investments and resources
• It will help the Education ministry to update and introduce new programs based on the
• prediction of job growth and workforce requirements

Data Collection
To forecast job in different industries in the UK we took data from - Office of National Statistics (GB), which contained time-series data of workforce employed in each sector.
The data source has quarterly data from 1978 of jobs in sectors such as construction, healthcare, manufacturing, education, retail and public administration.

If we visually analyze the data, we can see that there are some sectors where jobs have
grown and others where there is a decline in the number of jobs. Further, analysis
also shows that there is some seasonality in the quarterly data. Based on the analysis
in visualization software we can say there is both level and additive seasonality in
the data.

Final Forecasting Method:
From the exploratory analysis, we determined that there is trend and seasonality in the data. But when looked upon closely, we can determine that the seasonality component is not present to the same extents across all the sectors. While the Education sector shows the maximum seasonality, Public Admin and Defense sector shows the least seasonality present.

Therefore, since trend is present across all sectors and seasonality is present in varying degrees across all sectors, we have mostly employed Double Exponential and Holt Winter models to perform the final forecasting. The double exponential model captures the trend (but not seasonality) present in the data. And the Holt-Winter model captures both the trend and seasonality for a given time series.

After applying the Double Exponential model and Holt Winter model to all the sectors, we realized that Holt Winter gave a more accurate prediction compared to that of Double Exponential.

Since Holt Winter model gave the most accurate results, we can observe in the plot that the forecasted data fit very well with the actual data.

Thus, this model was reliably used to predict the numbers of jobs expected to be created across all the sectors for the next 5 years.

Error Metrics
To measure the performance of the 2 employed models, we primarily relied on the MAPE metric. It is also to be noted that for all the sectors there is no stark difference in the MAPE value of training data and that of validation data. This confirms that the forecasting model is not subject to the issue of overfitting.

Conclusion & Recommendation

• There is an increasing trend in job growth in the construction and human health sector
• Public administration and manufacturing sector has a stagnant growth in the workforce.
• There is a steady decline in the number of jobs being generated in the wholesale sector for the next 5 years.

Based on these findings we recommend the UK government to increase its spending in
the construction sector and the human healthcare sector. Also, the education ministry should increase seats in human health & social work-related activities and also increase intake in construction related courses based on the projection of the forecast for the next 5 years in these sectors.