Career options for graduate students in statistics?

Data mining (or machine learning) traditional statistics is relatively different. ?

Statistics: college curriculum includes multiple regression, serial correlation, econometrics, sample surveys, and more. Models built to focus on statistical tests and econometric tests. Software used SAS/SPSS/EVIEWS, models: regression models, principal components. The amount of data is not necessarily large. ?

Data mining: the model does not focus on statistical tests and econometric tests. Data volume is relatively large. Software: Python/R/SAS. model: random forest, logistic regression, SVM, neural network (there are also some deep learning models, CNN or something, alpha go used CNN), principal components. There should be a train set and a test set. Counting has many years of history, data mining is just more recently developed. I think data mining developed part of statistics and then added its own content. In the country to study statistics undergraduate will learn such as sample surveys and other courses that have nothing to do with data mining, and statistics experience so many years, teachers easily will not be on the curriculum will not make too much adjustment, so interested in data mining please read the related majors.

Traditional statistics modeling and sampling methods these are good, but I don't know about the employment situation now. I used to think that the employment is not too optimistic because the modeling of traditional statistics requires a PhD or a company that already has a mature model that is easily not modified. So I used to think that more statistics undergrads did descriptive analysis. Now I can't say that's still the case.

Recently I saw an undergraduate classmate (working for 3 years) posted a job posting for a data modeler, who should be familiar with data mining models, they use SAS/python/sql, and the salary range is 15K-30K, and the salary depends on the job match and experience. I suddenly feel that domestic statistics employment now seems to be quite good, as long as it is related to data mining, modeling, not just describing statistics that salary will be higher.