Streams of work (old comement)

[epogrebnyak]

13:04 07.12.2015

• datasets to be used in R and Excel for further analysis/modelling
• at git: (G0) maintainable/readable code + data structure (G1) enhance graphs (G2) make full coverage of KEP publication + brush dataset, (G3) make friendly API and (G4) clean data export
  (G5) data transformation (seasonality/detrending wth HP, index time series)
• propose key features for online interactive shell for KEP, similar to FRED

(G0) maintainable/readable code + data structure

• decide on packages (convert,parse,store + read/select + plot) or flat structure

(G1) enhance graphs

• no rigged upper end on png plots
• extend x axis to end of current year
• add text name to graph title

(G2) make full coverage of KEP publication + brush dataset

(G3) make friendly API

# + also tests as samples

get_var_list(freq=None)
"""Returns a list of (a)nnual, (q)tr and (m)onthly variables in database."""
{"a": [...], "q": [...], "m": [...], "all": [...]}

get_df(var_list, freq):
"""Returns dataframe with *var_list* variables"""

get_ts(var_name, freq):
"""Returns time series base on *var_name*"""

a_list = get_var_list(freq="a")
q_list = get_var_list(freq="q")
m_list = get_var_list(freq="m")

# dfa, dfq and dfm are complete contents of the database
dfa = get_df(a_list)
dfq = get_df(q_list)
dfm = get_df(m_list)

(G4) clean data export

• add NA to CSV/Excel export
• indicate which variables are a, q or m in 'variables sheet'

(G5) data transformation (seasonality/detrending wth HP, index time series)

• add seasonal component for selected variables (output monthly_nsa, monthly_sa, monthly_trend)
• add trend component for selected variables (output monthly_nsa, monthly_sa, monthly_trend)
• dating resession: data structure and background picture