An introductory example of using Wattsight API python library for searching, fetching, and plotting Wattsight data series. A number of use cases is presented in order to introduce different data structures.

Wattsight API python library
Jupyter (iPython) Notebooks
Pandas Data Analysis library
Plotly Charting library
World Timezone Definitions for Python

1. Using pip

   pip install -U wapi-python

Complete working code can be found in the attached Jupyter Notebook file.

1. Establish a session using client credentials.

   This requires having an OAuth client, which can be created at https://auth.volueinsight.com/account/oauth-clients.

   >>> session = wapi.Session(client_id='client id', client_secret='client secret')

2. Search curves by meta data.

   For this use case define meta criteria for DK1 area wind power production.

   >>> curves = session.search(commodity='POW',
                               category='WND',
                               unit='MWh/h',
                               area='DK1',
                               data_type=['A', 'F']
                              )
   >>> [c.name for c in curves]
   ['pro dk1 wnd intraday ec00da mwh/h cet min15 f',
    'pro dk1 wnd intraday lastec mwh/h cet min15 f',
    'pro dk1 wnd intraday tso mwh/h cet min15 f',
    'pro dk1 wnd ec00 mwh/h cet min15 f',
    'pro dk1 wnd ec12 mwh/h cet min15 f',
    'pro dk1 wnd gfs00 mwh/h cet min15 f',
    'pro dk1 wnd gfs06 mwh/h cet min15 f',
    'pro dk1 wnd gfs12 mwh/h cet min15 f',
    'pro dk1 wnd gfs18 mwh/h cet min15 f',
    'pro dk1 wnd ec00ens mwh/h cet min15 f',
    'pro dk1 wnd ec12ens mwh/h cet min15 f',
    'pro dk1 wnd mwh/h cet min15 a']

3. For the first example, filter on intraday curves and retrieve data for the selected curves.

   >>> intraday_curves = [c for c in curves if 'INTRADAY' in c.categories]
   >>> data = {}
   >>> label = {}
   >>> issue_date = pytz.timezone('CET').localize(datetime(2018, 4, 12))
   >>> date_from = issue_date + timedelta(minutes=15)
   >>> date_to = date_from + timedelta(hours=24)
   >>> for c in intraday_curves:
   >>>     if c.curve_type == 'INSTANCES':
   >>>         ts = c.get_instance(issue_date)
   >>>         if ts is not None:
   >>>             label[c.name] = ';'.join(c.sources)
   >>>             data[c.name] = ts.to_pandas()
   >>>     elif c.curve_type == 'TIME_SERIES':
   >>>         data[c.name] = c.get_data(data_from=date_from, 
   >>>                                   data_to=date_to).to_pandas()
   >>>         label[c.name] = 'Actual'

4. Plot selected data

   >>> fig = tls.make_subplots(len(data), cols=1, shared_xaxes=True)
   >>> for key in data.keys():
   >>>     ts = data[key]
   >>>     fig.append_trace({'x': ts.index, 'y': ts.values, 'type': 'scatter', 
   >>>                       'name': label[key]}, 1, 1)
   >>> title = ' '.join([c.commodity,
   >>>                   ' '.join(c.categories),
   >>>                   c.area,
   >>>                   issue_date.strftime("%Y-%m-%d %H:%M"),
   >>>                   c.time_zone])
   >>> fig['layout'].update(title=title)
   >>> fig['layout'].update(yaxis=dict(title=c.unit))    
   >>> offline.iplot(fig)

   

5. Retrive data for more instances and build contract development series.

   >>> curve = intraday_curves[0]
   >>> latest_instance = curve.get_latest(with_data=False)
   >>> issue_date = parser.parse(latest_instance.issue_date).astimezone(tz=latest_instance.tz)
   >>> value_date = issue_date + timedelta(minutes=15)
   >>> index = []
   >>> values = []
   >>> forecast_exists = True
   >>> while forecast_exists:
   >>>     ts = curve.get_instance(issue_date=issue_date)
   >>>     try:
   >>>         ts = curve.get_instance(issue_date=issue_date).to_pandas()
   >>>         values.append(ts[value_date])
   >>>         index.append(issue_date)
   >>>         issue_date = issue_date - timedelta(minutes=15)
   >>>     except:
   >>>         forecast_exists = False

   

6. For an example of working with tagged data, filter out weather data forecast curves.

   >>> curves_at00 = [c for c in curves 
                      if 'INTRADAY' not in c.categories 
                      and (c.data_type == 'A' or c.sources [0] in ["EC00", "EC00Ens", 'GFS00'] )
                     ]
   >>> [c.name for c in curves_at00]
   ['pro dk1 wnd ec00 mwh/h cet min15 f',
    'pro dk1 wnd gfs00 mwh/h cet min15 f',
    'pro dk1 wnd ec00ens mwh/h cet min15 f',
    'pro dk1 wnd mwh/h cet min15 a']

7. Retrive data for the selcted curves.

   >>> data = {}
   >>> label = {}
   >>> issue_date = pytz.timezone('CET').localize(datetime(2018, 4, 12))
   >>> date_to = date_from + timedelta(days=14)
   >>> for c in curves_at00:
   >>>     ts = None
   >>>     if c.curve_type == 'TAGGED_INSTANCES':
   >>>         ts = c.get_instance(issue_date=issue_date, tag='Avg')
   >>>         label[c.name] = ' '.join([';'.join(c.sources), 'Avg'])
   >>>     elif c.curve_type == 'INSTANCES':
   >>>         ts = c.get_instance(issue_date=issue_date)
   >>>         label[c.name] = ';'.join(c.sources)
   >>>     elif c.curve_type == 'TIME_SERIES':
   >>>         data[c.name] = c.get_data(data_from=date_from, data_to=date_to).to_pandas()
   >>>         label[c.name] = 'Actual'
   >>>     if ts is not None:
   >>>         data[c.name] = ts.to_pandas()