Supported metrics

Each metric requires a name, field and agg.

  • field is the field you are aggregating on.
  • agg is the type of aggregation you are going to perform.
  • name is an field to rename the output of the aggregation.

The metrics we support are:
avg (alias:mean), max, min, sum, count, cardinality, percentiles,
variance, std_deviation, std_deviation_bounds, kurtosis, sum_of_squares,
multifield metrics: correlation

avg, max, min, sum

These metrics work on fields with numeric type.

avg_price_metric = {"name": "avg_price", "field": "price", "agg": "avg"}

max_price_metric = {"name": "max_price", "field": "price", "agg": "max"}

min_price_metric = {"name": "min_price", "field": "price", "agg": "min"}

total_price_metric = {"name": "total_price", "field": "price", "agg": "sum"}

percentiles

Returns the different percentiles of the data (default: 1, 5, 25, 50, 75, 95, 99). This metric works on fields with numeric type.

percentiles_price_metric = {"field": "price", "agg": "percentiles"}

variance, std_deviation, std_deviation_bounds, sum_of_squares, kurtosis, skewness

We support a variety of advanced statistical measurements to help you do a further in-depth analysis of the data. These metrics work on fields with numeric type.

var_price_metric = {"field": "price", "agg": "variance"}

std_price_metric = {"field": "price", "agg": "std_deviation"}

std_bounds_price_metric = {"field": "price", "agg": "std_deviation_bounds"}

ss_price_metric = {"field": "price", "agg": "sum_of_squares"}

kurtosis_price_metric = {"field": "price", "agg": "kurtosis"}

skewness_price_metric = {"field": "skewness", "agg": "kurtosis"}

correlation

Correlation metric works on fields with numeric type, and requires at least 2 fields specified.

correlation_metric = {"fields": ["price", "revenue"], "agg": "correlation"}

count

Count returns the frequency of each groupby, it is returned by default.

suburb_count = {"name": "num_suburbs", "field": "propertyDetails.suburb", "agg": "count"}

cardinality

These metrics work on fields with text and numeric type. Returns the number of unique values in the field.

num_unique_suburbs = {"name": "num_suburbs", "field": "propertyDetails.suburb", "agg": "cardinality"}