Incorrect Historical Data

I observed some incorrect data while fetching historical data for Infy.

There's a spike as shown in the figure, which seems incorrect.
Here's the data,
[(1005.98, '2015-06-12T09:15:00+05:30'), (1008.43, '2015-06-12T09:20:00+05:30'), (1007.55, '2015-06-12T09:25:00+05:30'), (1007.78, '2015-06-12T09:30:00+05:30'), (1007.68, '2015-06-12T09:35:00+05:30'), (1007.78, '2015-06-12T09:40:00+05:30'), (1005.68, '2015-06-12T09:45:00+05:30'), (1006.3, '2015-06-12T09:50:00+05:30'), (1007.7, '2015-06-12T09:55:00+05:30'), (1008.9, '2015-06-12T10:00:00+05:30'), (1010.45, '2015-06-12T10:05:00+05:30'), (1010, '2015-06-12T10:10:00+05:30'), (1010.28, '2015-06-12T10:15:00+05:30'), (1011, '2015-06-12T10:20:00+05:30'), (1010.53, '2015-06-12T10:25:00+05:30'), (1011, '2015-06-12T10:30:00+05:30'), (1010.5, '2015-06-12T10:35:00+05:30'), (1010.98, '2015-06-12T10:40:00+05:30'), (1010.83, '2015-06-12T10:45:00+05:30'), (1010, '2015-06-12T10:50:00+05:30'), (1008.48, '2015-06-12T10:55:00+05:30'), (1008.5, '2015-06-12T11:00:00+05:30'), (1008.13, '2015-06-12T11:05:00+05:30'), (1007.65, '2015-06-12T11:10:00+05:30'), (1006.4, '2015-06-12T11:15:00+05:30'), (1006.75, '2015-06-12T11:20:00+05:30'), (1007.88, '2015-06-12T11:25:00+05:30'), (1007.25, '2015-06-12T11:30:00+05:30'), (1008.48, '2015-06-12T11:35:00+05:30'), (1007.75, '2015-06-12T11:40:00+05:30'), (1008.35, '2015-06-12T11:45:00+05:30'), (1007.83, '2015-06-12T11:50:00+05:30'), (1008.98, '2015-06-12T11:55:00+05:30'), (1009.38, '2015-06-12T12:00:00+05:30'), (1010.45, '2015-06-12T12:05:00+05:30'), (1009.88, '2015-06-12T12:10:00+05:30'), (1009.6, '2015-06-12T12:15:00+05:30'), (1000.7, '2015-06-12T12:20:00+05:30'), (1001.6, '2015-06-12T12:25:00+05:30'), (997.1, '2015-06-12T12:30:00+05:30'), (988.78, '2015-06-12T12:35:00+05:30'), (989.75, '2015-06-12T12:40:00+05:30'), (990, '2015-06-12T12:45:00+05:30'), (994.2, '2015-06-12T12:50:00+05:30'), (997.05, '2015-06-12T12:55:00+05:30'), (996, '2015-06-12T13:00:00+05:30'), (995, '2015-06-12T13:05:00+05:30'), (994.2, '2015-06-12T13:10:00+05:30'), (994.35, '2015-06-12T13:15:00+05:30'), (993.65, '2015-06-12T13:20:00+05:30'), (991.73, '2015-06-12T13:25:00+05:30'), (994.5, '2015-06-12T13:30:00+05:30'), (994.5, '2015-06-12T13:35:00+05:30'), (994.25, '2015-06-12T13:40:00+05:30'), (991.03, '2015-06-12T13:45:00+05:30'), (992.1, '2015-06-12T13:50:00+05:30'), (990.3, '2015-06-12T13:55:00+05:30'), (987.5, '2015-06-12T14:00:00+05:30'), (987.5, '2015-06-12T14:05:00+05:30'), (984.35, '2015-06-12T14:10:00+05:30'), (979.6, '2015-06-12T14:15:00+05:30'), (982, '2015-06-12T14:20:00+05:30'), (980.88, '2015-06-12T14:25:00+05:30'), (982.18, '2015-06-12T14:30:00+05:30'), (987.5, '2015-06-12T14:35:00+05:30'), (987.35, '2015-06-12T14:40:00+05:30'), (985.8, '2015-06-12T14:45:00+05:30'), (985.5, '2015-06-12T14:50:00+05:30'), (986, '2015-06-12T14:55:00+05:30'), (984.5, '2015-06-12T15:00:00+05:30'), (986.03, '2015-06-12T15:05:00+05:30'), (990.6, '2015-06-12T15:10:00+05:30'), (990.28, '2015-06-12T15:15:00+05:30'), (985, '2015-06-12T15:20:00+05:30'), (989, '2015-06-12T15:25:00+05:30')]
  • gauravmane
    There's a similar problem with other stocks too for 5min chart, so I assume the problem will be there for other time intervals too.
    For eg,


    And there're more stocks
    Adani, Ajanta Pharma, Arvind, Auropharma, Bata India, BEL, Cadila, Colpal, DHFL, Divi's labs, Federal Bank, Godfrey Philips, HCL Tech, IDFC, Infibeam, Kotak, Motherson, Repco Home
    Could you please look into this? There's seems to be a similar pattern of noise in all of these stocks.

    Also I observed that the splits/bonuses are not taken into consideration for 'day' data, for eg. ACC

    is it done deliberately?
  • sujith
    This is a minute level data what we have and has been reconciled with third-party vendors as well.
    If you are expecting a 100% accurate intra-day data then I would suggest getting it directly from the exchange.
  • sujith
    We haven't adjusted data for corporate actions prior to 2006. This could also be a corporate action. You can go to the exchange website to check if there was a corporate action on the same day.
  • gauravmane
    Okay, I am fine with the day chart not being adjusted for corporate action.
    But historical data for minute/5minute etc. seems to have noise like the spike shown in all above figures, which results into incorrect results on our end.
    Isn't there a way to rectify it on your end?
  • sujith
    As mentioned above, the dataset has already been reconciled. I would suggest getting it from the exchange.
Sign In or Register to comment.