Upload Demand Data

Upload your demand data as CSV, line-delimited JSON or JSON.

Beam Analyses support daily or weekly demand data. The type of data is detected automatically on upload.

For daily demand the minimum amount of data required is 6 months, or 180 data points. Beam can impute up to 20% of missing values, provided that the minimum data requirements are satisfied and there are no more than 7 consecutive days missing.

For weekly demand the minimum amount of data required is 2 years, or 104 data points. Beam can impute up to 10% of missing values, provided that the minimum data requirements are satisfied and there are no more than 5 consecutive weeks missing. All weekly data points must align with the start of the week and occur on the same weekday.

All data points need to be between 2017-01-01 and one year into the future from today's date (today + 365 days).

Uploading data replaces existing data for the same date. It's not currently possible to remove data for a particular date. The idea with this endpoint is that you're continuously adding new demand data over time.

The data can be uploaded in chunks, but the analysis data must remain valid at all times. If a newly uploaded chunk renders the analysis data invalid (for example, by introducing a large gap between data points), the chunk will be rejected, and the previously stored demand data will be preserved.

Request

HTTP Request

POST https://api.predicthq.com/v1/beam/analyses//sink

Path Parameters

Parameter	Description
analysis_id	An existing Beam Analysis ID.

Request Headers

Header	Value
Content-Type	Must be one of the following: text/csv application/x-ndjson application/json application/x-ldjson (deprecated) Note: application/x-ldjson has been deprecated in favor of application/x-ndjson. However the data format is exactly the same (newline delimited JSON). We recommend using application/x-ndjson

Request Body

You can upload the demand data for your analysis in any of the following formats:

CSV

The request body should contain comma-separated values that represent multiple data points with the column names (in lowercase) as date and demand, demonstrated by the example below:

date,demand
2023-01-01,12.235
2023-01-02,11.4
2023-01-03,12.14

The following request headers must be set:

Header	Value
Content-Type	text/csv

Column Types:

Column	Description
date string required	ISO8601 Date format (YYYY-MM-DD) E.g., 2023-01-01
demand number required	Number value (float or integer, must be a positive number). Demand will typically be the demand you use in your demand forecast if you are forecasting. For example, it could be units sold, room bookings, or number of staff rostered on per day or any other unit. E.g., 12.235

Newline-delimited JSON

The request body should consist of line-delimited JSON objects, each representing a distinct data point. Ensure that the fields within these objects are named date and demand (lowercase). The format should follow the example given below:

{"date": "2023-01-01", "demand": 12.235}
{"date": "2023-01-02", "demand": 11.4}
{"date": "2023-01-03", "demand": 12.14}

The following request headers must be set:

Header	Value
Content-Type	application/x-ndjson

JSON Fields:

Field	Description
date string required	ISO8601 Date format (YYYY-MM-DD) E.g., 2023-01-01
demand string required	Number value (float or integer, must be a positive number). Demand will typically be the demand you use in your demand forecast if you are forecasting. For example, it could be units sold, room bookings, or number of staff rostered on per day or any other unit. E.g., 12.235

JSON

Note that this format is only suitable for appending data to an already valid data set.

The request body should contain a JSON object that represents a single data point. For each request, only one data point is permitted. Ensure that the fields are named date and demand (lowercase), as demonstrated in the example below:

{
  "date": "2023-01-01",
  "demand": 12.235
}

The following request headers must be set:

Header	Value
Content-Type	application/json

JSON Fields:

Field	Description
date string required	ISO8601 Date format (YYYY-MM-DD) E.g., 2023-01-01
demand string required	Number value (float or integer, must be a positive number). Demand will typically be the demand you use in your demand forecast if you are forecasting. For example, it could be units sold, room bookings, or number of staff rostered on per day or any other unit. E.g., 12.235

Response

Success Response

If successful, the HTTP response code will be 202 Accepted.

Error Response

An unsuccessful HTTP response code could be returned for several reasons. In addition to an error message, there may also be a code field when applicable. The table below outlines the meaning of several error codes that may be returned.

Code	Description
content_type_invalid	The Content-Type header is unsupported.
data_validation_failed	The format of the data uploaded is incorrect.
json_no_data	An empty JSON body was uploaded.
json_invalid_format	The uploaded JSON is formatted incorrectly. Please ensure JSON data is correctly UTF-8 encoded and that there are no invalid escape sequences.
ndjson_no_data	An empty NDJSON body was uploaded.
ndjson_invalid_format	The NDJSON body is formatted incorrectly. Please ensure NDJSON data is correctly UTF-8 encoded and that there are no invalid escape sequences.
csv_invalid_row	The uploaded CSV has an invalid row.
csv_invalid_header	The uploaded CSV headers are incorrect. Please use date,demand.
csv_no_data	The uploaded CSV is empty or only has headers set.
csv_invalid_format	The uploaded CSV is formatted incorrectly. Please ensure CSV data is correctly UTF-8 encoded and that there are no invalid escape sequences.
start_date_invalid	The earliest date in the uploaded demand data is before 2017-01-01.
end_date_invalid	The latest date in the uploaded demand data is more than 1 year into the future.
duplicate_rows	The uploaded demand data contains duplicate dates. Please remove all duplicates before uploading.
no_data	There is no demand data.
weekly_demand_date_check_failed	The weekly demand dates do not start on the same weekday.
below_minimum_threshold	There are not enough data points in the uploaded demand data.
constant_demand_column	The demand column is constant.
exceeding_missing_data_percentage	There are too many missing demand values.
exceeding_consecutive_missing_data_volume	There are too many consecutive missing demand values.
high_variance_difference_rate	The variance of the uploaded demand data is too large to be processed.
unexpected_error	An unexpected error occurred.

Examples

curl

curl -X POST "https://api.predicthq.com/v1/beam/analyses/$ANALYSIS_ID/sink" \
     -H "Content-Type: text/csv" \
     -H "Authorization: Bearer $ACCESS_TOKEN" \
     --data @data.csv

python

import requests

response = requests.post(
    url="https://api.predicthq.com/v1/beam/analyses/$ANALYSIS_ID/sink",
    headers={
        "Authorization": "Bearer $ACCESS_TOKEN",
        "Content-Type": "text/csv"
    },
    data=open("data.csv")
)

print(response.status_code)

Guides

Below are some guides relevant to this API:

• Beam Guides