Technical Documentation | Glossary

This following words and concepts are defined in the context of the Metamarkets analytics platform.

Words that appears in italics are also defined in this glossary.

Backfill or Batch Clean Up Adding missing data to a data set that has already been ingested by the Metamarkets Realtime Data Ingestion (RDI) platform.

Batch Files The files containing the data being uploaded to the Metamarkets HTTP endpoints. Also called "data files" and log files.

Cardinality Used to describe the degree to which dimensions aggregate in druid. Dimensions with high cardinality will aggregate poorly, for example zip code.

Dashboard The Metamarkets dashboard (or "dash") is a web-based analytics application used to view and query data in real time.

Data Ingestion The process of taking in data that is being uploaded. Metamarkets Realtime Data Ingestion (RDI) is the platform that ingests data via an HTTP endpoint, process it, and then surfaces data in the Metamarkets dashboard.

Dimension A field containing values that are typically strings and are not aggregated (e.g., country codes). Dimensions with values that can be aggregated are usually treated as metrics.

Event A record summarizing a transaction, such as results from an auction. Events are encapsulated as JSON objects.

HTTP endpoint An interface with a URL for an address. The URL is used by the remote program to connect to the endpoint. The endpoints provided by Metamarkets require HTTPS (encrypted HTTP) to protect credentials and data.

Log Files The files containing the data being uploaded to the Metamarkets HTTP endpoints. Also called "data files" and batch files.

Lookup Table A file containing key-value mappings. Typically a code (the key) is mapped to a longer string of characters (the value). This allows senders to create more compact data since only the codes need to be provided. After the data is ingested, the lookup table is used to expand the data with the stored values.

Metric A field containing values that can be aggregated, such as number of clicks or impressions served. Data metrics are contained in or inferred from raw data. Computed metrics are calculated after data is ingested and while it is being processed.

No Bid A valid bid response (status of 1 in the bid response object) that contains an empty seat bid.

OpenRTB API Specification The Open Real-Time Bidding API specification is the main standard on which the schema required for using the Metamarkets Realtime Data Ingestion (RDI) is based.

Realtime Data Ingestion (RDI) Realtime Data Ingestion is the Metamarkets data intake and processing platform. RDI is designed to ingest, process, and surface current data in seconds.

Record An event that summarizes a single time-series data point, encapsulated in a JSON object. An auction record, for example, summarizes an RTB auction, including bids.

S3 Simple Storage Service is on-line file storage, part of Amazon Web Services. The root S3 storage partition is referred to as a bucket.