-) Create dataset
employment_city_sector.csv: sectoral employment share of total employment
* 3 sectors: agriculture, manufacturing, services
* Define "services" to be: retail sector, the financial sector, business adminis- tration, leisure and cultural activities
- gdp_pop_city.xlsx: gdp, population
2) Regression analysis
Regress by OLS sectoral employment share on log gdp per capita
- Do for each of the three sectors
- Use robust standard errors
3) Create graphs like the employment one in figure 1
-y-variable: sectoral employment share of total employment
x-variable: log gdp per capita
- Each point will be a city rather than a country, in a single year
- Add line of best fit with the coefficient on sectoral employment share (this is found in part 2)
* Write coefficient as coefficient-magnitude (coefficient standard error), round each to 3 decimal places
* This code will be helpful: local coeff: di %3.2f
- Label axes (not with variable name e.g. loggdp but Log GDP), add graph title
(sector name)
- x-axes ticks should have e.g. 10$^2$ not 2.
- i.e Make the graph look nice and readable.
