🛠️Setup your Analysis

Covers how to configure and prepare data for analysis.

Set Column Meta

The "Set Column Meta" method allows you to define metadata for the columns that are present in your dataset. By setting column metadata, you ensure that the library correctly interprets and processes your data during analysis.

Method Signature:

set_column_meta(
    date_column: Union[dict, str] = None, 
    metric_column: Union[list, dict, str] = None,
    dimension_column: Union[list, str] = None
) -> None

Parameter

Type

Description

date_column

str/dict

Specify name of the column in your dataset that contains date related information. Essential for accurate analysis and enabling the library to recognize and treat date values appropriately. For more details refer here. Default: None

metric_column

str or List[str/dict]

Specify the name of the column/columns that holds numerical data. This parameter helps identify the metrics that will be used for analysis, facilitating data-driven insights. You can pass a single column name using string and multiple columns using a list. For more details refer here. Default: None

dimension_column

List[str]

Specify the name of column/columns that contains categorical data. It defines the dimensions, categories, or groupings within your dataset.

You can pass a single column name using string and multiple columns using a list. Default: None

# Example of setting column metadata
phrzr.set_column_meta(
    date_column="Date",
    metric_column=["Sales", "Quantity"],
    dimension_column=["Region", "State"]
)

In the above example, we've configured the metadata for columns as follows:

"Date" is the date column with a string data type.
"Sales" and "Quantity" are metric columns with integer data types.
"Region" and "State" are the dimensions to be analysed.

Set Analysis Period

To enhance the flexibility of data analysis users can specify the period aggregation for a particular date column within their dataset. This empowers users to tailor the granularity of their data analysis according to their specific needs.

To define period for your analysis, you will have to pass dictionary as a value in the date_column parameter of set_column_meta method with the following keys:

Key

Description

name

Name of the Date Column

period

Granular level at which you want to analyse your data. Accepted Values - "day", "week", "month", "quarter", "year"

If a date column is provided as a string, the assumption is that the period is set to month by default.

Example

To analyse your data at Weekly level you can set your column meta as follow:

phrzr.set_column_meta(
    date_column={"name": "Date", "period": "week"},
    metric_column="Sales",
)

Set Metric Aggregation

Provides users with the capability to define and customize the aggregation method applied to a specific metric within their dataset. This functionality empowers users to choose how their data is summarized and analyzed, tailoring the analysis to meet their specific analytical requirements.

To define aggregation for your metric you will have to pass list of dictionaries as value in the metric_column parameter of set_column_meta method with the following keys:

Key

Description

name

Name of the Date Column

aggregation

Aggregation to apply on metric Accepted Values - "sum", "avg", "min", "max", "count"

If a metric is provided without an explicit aggregation key, the default behavior is to apply the sum aggregation method.

Example

You can analyse the Avg Weekly Sales using the following snippet:

phrzr.set_column_meta(
    date_column={"name": "Date", "period": "week"},
    metric_column=[{"name": "Sales", "aggregation": "avg"}]
)

You can also set individual aggregation for each metric.

phrzr.set_column_meta(
    date_column={"name": "Date", "period": "week"},
    metric_column=[
            {"name": "Sales", "aggregation": "avg"},
            {"name": "Profit", "aggregation": "sum"},
            {"name": "Product", "aggregation": "count"}
        ]
)

Choose your Analysis

Following are the various types of analyses that can be used to generate insights/summaries for your data. These analyses help you gain deeper insights and make informed decisions. Here's an overview of the different types of analyses available: