Intelligent data ingestion

A method, comprising:

• receiving, by one or more processors coupled with memory, via a network, a data file comprising a plurality of columns and a plurality of rows;
• detecting, by the one or more processors, that a first column of the plurality of columns in the data file lacks a column header name;
• determining, by the one or more processors, an adjacency context for the first column that includes, for each of one or more columns adjacent to the first column, (i) a respective column header name and (ii) a descriptor of values in the one or more columns adjacent to the first column;
• generating, by the one or more processors, with a proximity-based machine learning model, a target column name for the first column based on the adjacency context;
• and
• ingesting, by the one or more processors, values of the first column into a database using the target column name as the column header name.

A non-transitory computer-readable medium storing processor-executable instructions that, when executed by one or more processors, cause the one or more processors to:

• receive, via a network, a data file comprising a plurality of columns and a plurality of rows;
• detect that a first column of the plurality of columns in the data file lacks a column header name;
• determine an adjacency context for the first column that includes, for each of one or more columns adjacent to the first column, (i) a respective column header name and (ii) a descriptor of values in the one or more columns adjacent to the first column;
• generate, with a proximity-based machine learning model, a target column name for the first column based on the adjacency context;
• and
• ingest values of the first column into a database using the target column name as the column header name.