Spécifications
I need python script to download lastest macroeconomical indicators from official statistic bureaus or central banks.
Sources:
1. Central banks: feed, european central bank, snb, rba, rbnz, all wich you can access.
2. Statistic bureaus: eurostat, cabinet offices, statistic bureaus, all wich you can acces.
3. research sites for cenral banks, governments, for forecasting, ism, universities (like michigan), all what you are abile.
not 4. trid party web providers or databasses: not desired becose of missing or not updated issues, they acteptable only as last option... quandl, ceic, tradingeconomic...
Your focus:
1. download indicator for free with or without free api key
2. create raw dataframe in columns date, obseved value (values) no modification, no manipulation, no resampling, no grouping, just simple dataframe with original data transormed to simple vertical table
3. build additional subfunction: resampling, groupby (month, week) for resampling (if two or more observations in one targeted time/date period)
4. create jupyter notebook for every function separatly
5. solving web page ban if possible: some webpages do ban you if you downloading to much or often from they web page example census.gov
6. sql table integration: updating old values if source provide new values in old rows: sometimes this month valu will be changed next month so we need lastest actual: target mariadb or postsql or some free (for business) mysql-like solution
7. optimization of speed low internet trafic, targeting one or multiple indicators at once (one reqest downloading more indicators if possible - savig time and internet traffic)
8. research of datasources, checking if values same like original source, searching for free databases only (we do not pay for data)
9. preparing data for forecasting by Machine Learning: shift dataframe, handling nan (ffill, mode, mean,...). creating testing / Researching dataset
Your profil
1. newbee you will learn how to do this
2. amateur you are hobbiest you did something of this
3. you somebody wich triing develop own dataset for machine learning
4. you dataminer, machine learning engenieer, datascientist
About project:
1. it is private and propetary
2. the code will not be opensource becose of this I need your work to be simple and minimalistic so you have no conflict of interrest, I have no problem if you using your code or my concepsts for your private activity...
3. my own licence
About me
1. hobby programator: python mql c like languages
2. hobby trader: funamental, sentimental, technical, quantitative, machinelearning