Method and system for manipulating weather forecasts

A non-transitory, machine-readable medium, comprising instructions, which when performed by a processor of a machine, causes the processor to perform operations to:

• provide first training data to a machine learning model, the first training data relating to predicting a first weather forecast using first weather-related datapoints;
• train the machine learning model with the first training data;
• receive a set of second weather-related datapoints from a plurality of sources;
• cause a first user interface to be displayed, the first user interface including:
• a list of the set of second weather-related datapoints that lists the plurality of sources associated with each of the second weather-related datapoints;
• and
• a plurality of inputs at each datapoint of the set of second weather-related datapoints;
• receive user inputs at ones of the plurality of inputs, wherein:
• first ones of the user inputs correspond to weights relating to each source of the plurality of sources associated with each of the second weather-related datapoints;
• and
• second ones of the user inputs correspond to selected weather-related datapoints of the second weather-related datapoints;
• bias correct a subset of the set of second weather-related datapoints;
• generate, using the trained machine learning model, an estimated weather forecast for a location using the bias corrected subset second weather-related datapoints based on the weights associated with each source of the plurality of sources;
• determine an actual weather forecast at the location after the estimated weather is determined;
• create second training data using the actual weather forecast;
• provide the second training data to the machine learning model;
• and
• re-train the trained machine learning model with the second training data such that machine learning model changes over time.

A method comprising:

• providing first training data to a machine learning model, the first training data relating to predicting a first weather forecast using first weather-related datapoints;
• training the machine learning model with the first training data;
• receiving a set of second weather-related datapoints from a plurality of sources;
• causing a first user interface to be displayed, the first user interface including:
• a list of the set of second weather-related datapoints that lists the plurality of sources associated with each of the second weather-related datapoints;
• and
• a plurality of inputs at each datapoint of the set of second weather-related datapoints;
• receiving user inputs at ones of the plurality of inputs, wherein:
• first ones of the user inputs correspond to weights relating to each source of the plurality of sources associated with each of the second weather-related datapoints;
• and
• second ones of the user inputs correspond to selected weather-related datapoints of the second weather-related datapoints;
• bias correcting a subset of the set of second weather-related datapoints;
• generating, using the trained machine learning model, an estimated weather forecast for a location using the bias corrected subset second weather-related datapoints based on the weights associated with each source of the plurality of sources;
• determining an actual weather forecast at the location after the estimated weather is determined;
• creating second training data using the actual weather forecast;
• providing the second training data to the machine learning model;
• and
• re-training the trained machine learning model with the second training data such that machine learning model changes over time.