Core Reporting API - Developer Guide

This document explains how to use the Core Reporting API to access Google Analytics Data.

Introduction

The Core Reporting API provides access to the tabular data in Google Analytics standard and custom reports. To access data, you create a query that specifies: the view (profile), the start and end dates, and the dimensions and metrics that make up the column headers in the table. This query is sent to the Core Reporting API and The Core Reporting API returns all the data in the form of a table.

If you are new to the API, read the Core Reporting API Overview for an introduction to purpose of the Core Reporting API and the data it provides.

Before You Begin

This guide demonstrates how to access the Google Analytics API using the Java, Python, PHP, and JavaScript programming languages.

  • Read the client libraries page for a complete list of programming language specific client libraries that work with the API.
  • Read the Reference Guide to access the API without a client library.

Each client library provides a single analytics service object to access all Core Reporting API data. To create the service object you generally have to go through the following steps:

  1. Register your application in the Google API Console.
  2. Authorize access to Google Analytics data.
  3. Create an Analytics service object.

If you haven't completed these steps, please stop and read the Hello Google Analytics API Tutorial. This tutorial will walk you through the initial steps of building a Google Analytics API application. Once complete, you will be able to use this guide to perform real-world tasks.

The following code snippet contains a variable to store an authorized service object.

Java

Analytics analytics = // Read Hello Analytics Tutorial for details.

Python

analytics = # Read Hello Analytics Tutorial for details.

PHP

$client = // Read Hello Analytics Tutorial for details.

// Return results as objects.
$client->setUseObjects(true);

$analytics = new apiAnalyticsService($client);

The PHP library will return all the API results as an associative array. To return real objects instead, you can call the client useObject method as demonstrated in the example above.

JavaScript

<script src="https://apis.google.com/js/client.js?onload=loadLib"</script>
<script>
function loadLib() {
  // Handle all the authorization work.
  // Read Hello Analytics Tutorial for details.
  gapi.client.load('analytics', 'v3', makeApiCall);
}
</script>

The first script tag loads the Google API JavaScript library. Once loaded, loadLib is executed to load the analytics service class. Once complete, the object gapi.client.analytics should exist in the DOM and be ready to use to query the Core Reporting API.

Once you create an analytics service object, you are ready to make requests to the Core Reporting API.

Note: The analytics service object can also be used to access the Management API.

Overview

An application that uses the Core Reporting API will generally follow 2 steps:

  • Query the Core Reporting API
  • Work with the API results

Let's look at both steps.

Query the Core Reporting API

Build a Core Reporting API query

The analytics service object contains a method to build a Core Reporting API query.

Each Core Reporting API query contains a set of parameters which specify what data to return.

One of the most important query parameters is the ids parameter, or the table ID. This parameter specifies from which Google Analytics view (profile) to retrieve data. The value is in the format ga:xxx where xxx is the view (profile) ID.

Java

Get apiQuery = analytics.data().ga()
    .get(tableId,                  // Table Id.
        "2012-01-01",              // Start date.
        "2012-01-15",              // End date.
        "ga:sessions")               // Metrics.
    .setDimensions("ga:source,ga:keyword")
    .setSort("-ga:sessions,ga:source")
    .setFilters("ga:medium==organic")
    .setMaxResults(25);

Python

api_query = service.data().ga().get(
    ids=TABLE_ID,
    start_date='2012-01-01',
    end_date='2012-01-15',
    metrics='ga:sessions',
    dimensions='ga:source,ga:keyword',
    sort='-ga:sessions,ga:source',
    filters='ga:medium==organic',
    max_results='25')

PHP

private function queryCoreReportingApi() {
  $optParams = array(
      'dimensions' => 'ga:source,ga:keyword',
      'sort' => '-ga:sessions,ga:source',
      'filters' => 'ga:medium==organic',
      'max-results' => '25');

  return $service->data_ga->get(
      TABLE_ID,
      '2010-01-01',
      '2010-01-15',
      'ga:sessions',
      $optParams);
}

In this library, the get method not only creates a Core Reporting API query, but also executes the request to the API.

JavaScript

function makeApiCall() {
  var apiQuery = gapi.client.analytics.data.ga.get({
    'ids': TABLE_ID,
    'start-date': '2010-01-01',
    'end-date': '2010-01-15',
    'metrics': 'ga:sessions',
    'dimensions': 'ga:source,ga:keyword',
    'sort': '-ga:sessions,ga:source',
    'filters': 'ga:medium==organic',
    'max-results': 25
  });
  // ...
}

In this example, the makeApiCall function is called once the JavaScript client library has loaded. Inside, the function creates a new Google Analytics API query and stores the object in the apiQuery variable.

A complete listing of all the query parameters and what they do can be found in the Core Reporting API reference guide. Also the dimension and metric parameters allow you to specify what data to retrieve from Google Analytics. A complete list can be found on the dimensions and metrics reference page.

Making a Core Reporting API data request

Once you have a query defined, you call it's execute method to send the query to Google Analytics servers.

Java

try {
  apiQuery.execute();
  // Success. Do something cool!

} catch (GoogleJsonResponseException e) {
  // Catch API specific errors.
  handleApiError(e);

} catch (IOException e) {
  // Catch general parsing network errors.
  e.printStackTrace();
}

If you prefer to access the raw API response instead, use the executeUnparsed() method:

HttpResponse response = apiQuery.executeUnparsed();

Python

try:
  results = get_api_query(service).execute()
  print_results(results)

except TypeError, error:
  # Handle errors in constructing a query.
  print ('There was an error in constructing your query : %s' % error)

except HttpError, error:
  # Handle API service errors.
  print ('There was an API error : %s : %s' %
         (error.resp.status, error._get_reason()))

PHP

  try {
    $results = queryCoreReportingApi();
    // Success. Do something cool!

  } catch (apiServiceException $e) {
    // Handle API service exceptions.
    $error = $e->getMessage();
  }

JavaScript

function makeApiCall() {
  // ...

  apiQuery.execute(handleCoreReportingResults);
}

function handleCoreReportingResults(results) {
  if (!results.error) {
    // Success. Do something cool!
  } else {
    alert('There was an error: ' + results.message);
  }
}

This example follows from the previous step where a Core Reporting API query was created. In this step, the query is executed. The parameter to the execute method is a reference to a callback function that will get executed once data is returned from the API.

Once the API returns with results, the callback function is executed and is passed the data from the API. If an error occurs, the results will contain a property named error.

In this example, a check is made to see if error exists or whether the API returned successfully.

If the query was successful, the API will return the requested data. If any errors occur, the API will return a specific status code and a message describing the error. All applications should properly catch and handle errors.

Working with the API results

If the Core Reporting API query was successful, the API returns with the Analytics reporting data as well as other related information about the data.

Analytics reporting data

The main data returned from the API can be thought of as a table with 2 main types of data:

  • The header that describes the types of values in each column
  • The rows of data in the table

Column header data

Each API response contains a column header field that represents the header information of the table. The field is a list (or an array) of objects where each object describes the type of data in the column. The column order is dimensions columns followed by metric columns in the same order as specified in the original query.

Java

private void printColumnHeaders(GaData gaData) {
 System.out.println("Column Headers:");

 for (GaDataColumnHeaders header : gaData.getColumnHeaders()) {
   System.out.println("Column Name: " + header.getName());
   System.out.println("Column Type: " + header.getColumnType());
   System.out.println("Column Data Type: " + header.getDataType());
 }
}

Python

def print_column_headers():
  headers = results.get('columnHeaders')

  for header in headers:
    # Print Dimension or Metric name.
    print 'Column name = %s' % header.get('name'))
    print 'Column Type = %s' % header.get('columnType')
    print 'Column Data Type = %s' % header.get('dataType')

PHP

private function printColumnHeaders(&results) {
  $html = '';
  $headers = $results->getColumnHeaders();

  foreach ($headers as $header) {
    $html .= <<<HTML
Column Name = {$header->getName()}
Column Type = {$header->getColumnType()}
Column Data Type = {$header->getDataType()}
HTML;

  print $html;
}

JavaScript

function printColumnHeaders(results) {
  var output = [];

  for (var i = 0, header; header = results.columnHeaders[i]; ++i) {
    output.push(
        'Name        = ', header.name, '\n',
        'Column Type = ', header.columnType, '\n',
        'Data Type   = ', header.dataType, '\n'
    );
  }

  alert(output.join(''));
}

Row data

The main data returned from the API is returned as a 2-dimensional List of strings. The outer list represents all the rows of data. Each inner list represents a single row, where the order of cells in a row is same as the fields in the column header object described above.

Since data in each cell is returned as a string, the DataType field in each column header object is particularly useful since it can be used to parse string values into an appropriate type. See the metadata API response for all the possible data types.

The following examples print both the headers and rows of the table.

Java

private void printDataTable(GaData gaData) {
 if (gaData.getTotalResults() > 0) {
   System.out.println("Data Table:");

   // Print the column names.
   for (GaDataColumnHeaders header : gaData.getColumnHeaders()) {
     System.out.format("%-32s", header.getName() + '(' + header.getDataType() + ')');
   }
   System.out.println();

   // Print the rows of data.
   for (List<String> rowValues : gaData.getRows()) {
     for (String value : rowValues) {
       System.out.format("%-32s", value);
     }
     System.out.println();
   }
 } else {
   System.out.println("No Results Found");
 }

Python

def print_data_table(results):
  # Print headers.
  output = []
  for header in results.get('columnHeaders'):
    output.append('%30s' % header.get('name'))
  print ''.join(output)

  # Print rows.
  if results.get('rows', []):
    for row in results.get('rows'):
      output = []
      for cell in row:
        output.append('%30s' % cell)
      print ''.join(output)
  else:
    print 'No Results Found'

PHP

private function printDataTable(&$results) {
  if (count($results->getRows()) > 0) {
    $table .= '<table>';

    // Print headers.
    $table .= '<tr>';

    foreach ($results->getColumnHeaders() as $header) {
      $table .= '<th>' . $header->name . '</th>';
    }
    $table .= '</tr>';

    // Print table rows.
    foreach ($results->getRows() as $row) {
      $table .= '<tr>';
        foreach ($row as $cell) {
          $table .= '<td>'
                 . htmlspecialchars($cell, ENT_NOQUOTES)
                 . '</td>';
        }
      $table .= '</tr>';
    }
    $table .= '</table>';

  } else {
    $table .= '<p>No Results Found.</p>';
  }
  print $table;
}

JavaScript

function printRows(results) {
  output = [];

  if (results.rows && results.rows.length) {
    var table = ['<table>'];

    // Put headers in table.
    table.push('<tr>');
    for (var i = 0, header; header = results.columnHeaders[i]; ++i) {
      table.push('<th>', header.name, '</th>');
    }
    table.push('</tr>');

    // Put cells in table.
    for (var i = 0, row; row = results.rows[i]; ++i) {
      table.push('<tr><td>', row.join('</td><td>'), '</td></tr>');
    }
    table.push('</table>');

    output.push(table.join(''));
  } else {
    output.push('<p>No Results Found</p>');
  }

  alert(output.join(''));
}

Report information

Along with the main table data, the data returned from the API contains some high level information about the response. You can print it using:

Java

private void printResponseInfo(GaData gaData) {
  System.out.println("Contains Sampled Data: " + gaData.getContainsSampledData());
  System.out.println("Kind: " + gaData.getKind());
  System.out.println("ID:" + gaData.getId());
  System.out.println("Self link: " + gaData.getSelfLink());
}

Python

def print_response_info(results):
  print 'Contains Sampled Data = %s' % results.get('containsSampledData')
  print 'Kind                  = %s' % results.get('kind')
  print 'ID                    = %s' % results.get('id')
  print 'Self Link             = %s' % results.get('selfLink')

PHP

private function printReportInfo(&$results) {
  $html = <<<HTML
  <pre>
Contains Sampled Data = {$results->getContainsSampledData()}
Kind                  = {$results->getKind()}
ID                    = {$results->getId()}
Self Link             = {$results->getSelfLink()}
</pre>
HTML;

  print $html;
}

JavaScript

function printReportInfo(results) {
  var output = [];

  output.push(
      'Contains Sampled Data  = ', results.containsSampledData, '\n',
      'Kind                   = ', results.kind, '\n',
      'ID                     = ', results.id, '\n',
      'Self Link              = ', results.selfLink, '\n');

  alert(output.join(''));
}

The containsSampledData field is important because it describes whether the API response has been sampled. Sampling can affect the results of your data and a common reason why the values returned from the API do not match the web interface. See the Sampling concept guide for more details.

View (Profile) information

Each response contains a group of parameters which indicate the Account, Web Property, and View (Profile) this data belongs to.

Java

private void printProfileInfo(GaData gaData) {
  GaDataProfileInfo profileInfo = gaData.getProfileInfo();

  System.out.println("Account ID: " + profileInfo.getAccountId());
  System.out.println("Web Property ID: " + profileInfo.getWebPropertyId());
  System.out.println("Internal Web Property ID: " + profileInfo.getInternalWebPropertyId());
  System.out.println("View (Profile) ID: " + profileInfo.getProfileId());
  System.out.println("View (Profile) Name: " + profileInfo.getProfileName());
  System.out.println("Table ID: " + profileInfo.getTableId());
}

Python

def print_profile_info(result):

  info = results.get('profileInfo')
  print 'Account Id          = %s' % info.get('accountId')
  print 'Web Property Id     = %s' % info.get('webPropertyId')
  print 'Web Property Id     = %s' % info.get('internalWebPropertyId')
  print 'View (Profile) Id   = %s' % info.get('profileId')
  print 'Table Id            = %s' % info.get('tableId')
  print 'View (Profile) Name = %s' % info.get('profileName')

PHP

private function printProfileInformation(&$results) {
  $profileInfo = $results->getProfileInfo();

  $html = <<<HTML
<pre>
Account ID               = {$profileInfo->getAccountId()}
Web Property ID          = {$profileInfo->getWebPropertyId()}
Internal Web Property ID = {$profileInfo->getInternalWebPropertyId()}
Profile ID               = {$profileInfo->getProfileId()}
Table ID                 = {$profileInfo->getTableId()}
Profile Name             = {$profileInfo->getProfileName()}
</pre>
HTML;

  print $html;
}

JavaScript

function printProfileInfo(results) {
  var output = [];

  var info = results.profileInfo;
  output.push(

      'Account Id          = ', info.accountId, '\n',
      'Web Property Id     = ', info.webPropertyId, '\n',
      'View (Profile) Id   = ', info.profileId, '\n',
      'Table Id            = ', info.tableId, '\n',
      'View (Profile) Name = ', info.profileName);

  alert(output.join(''));
}

Each of these ID corresponds to various entities in the Management API hierarchy. You can use these IDs to form Management API queries to get additional configuration information about the view (profile). For example, you can query the Management API Goals Collection to see which Goals are active along with their configured goal names.

Query information

Each Core Reporting API response contains an object that contains all the query parameter values used to create the response.

Java

private void printQueryInfo(GaData gaData) {
  GaDataQuery query = gaData.getQuery();

  System.out.println("Ids: " + query.getIds());
  System.out.println("Start Date: " + query.getStartDate());
  System.out.println("End Date: " + query.getEndDate());
  System.out.println("Metrics: " + query.getMetrics()); // List
  System.out.println("Dimensions: " + query.getDimensions());
  System.out.println("Sort: " + query.getSort()); // List
  System.out.println("Segment: " + query.getSegment());
  System.out.println("Filters: " + query.getFilters());
  System.out.println("Start Index: " + query.getStartIndex());
  System.out.println("Max Results: " + query.getMaxResults());
}

Python

def print_query_info(results):
  query = results.get('query')
  for key, value in query.iteritems():
    print '%s = %s' % (key, value)

PHP

private function printQueryParameters(&$results) {
  $query = $results->getQuery();

  $html = '<pre>';
  foreach ($query as $paramName => $value) {
    $html .= "$paramName = $value\n";
  }
  $html .= '</pre>';

  print $html;
}

JavaScript

function printQuery(results) {
  output = [];

  for (var key in results.query) {
    output.push(key, ' = ', results.query[key], '\n');
  }

  alert(output.join(''));
}

The metrics and sort parameters are returned as values in a list, while the other parameters are returned as strings.

Pagination information

Any Core Reporting API request might match hundreds of thousands of rows of Google Analytics data. The Core Reporting API will only return a subset at a given time, which can be referred to as a single page of data. You use the pagination fields to retrieve all pages of data.

Java

private void printPaginationInfo(GaData gaData) {
  System.out.println("Items Per Page: " + gaData.getItemsPerPage());
  System.out.println("Total Results: " + gaData.getTotalResults());
  System.out.println("Previous Link: " + gaData.getPreviousLink());
  System.out.println("Next Link: " + gaData.getNextLink());
}

Python

def print_pagination_info(results):
  print 'Items per page = %s' % results.get('itemsPerPage')
  print 'Total Results  = %s' % results.get('totalResults')
  print 'Previous Link  = %s' % results.get('previousLink')
  print 'Next Link      = %s' % results.get('nextLink')

PHP

private function getPaginationInfo(&$results) {
  $html = <<<HTML
<pre>
Items per page = {$results->getItemsPerPage()}
Total results  = {$results->getTotalResults()}
Previous Link  = {$results->getPreviousLink()}
Next Link      = {$results->getNextLink()}
</pre>
HTML;

  print $html;
}

JavaScript

function printPaginationInfo(results) {
  var output = [];

  output.push(
      'Items Per Page = ', results.itemsPerPage, '\n',
      'Total Results  = ', results.totalResults, '\n',
      'Previous Link  = ', results.previousLink, '\n',
      'Next Link      = ', results.nextLink, '\n');

  alert(output.join(''));
}

The totalResults field represents the total number of rows of data that your query matched in Google Analytics. This can be greater than the actual number of rows returned in a single page of the response. The itemsPerPage field represents the number of rows returned in this page.

The previousLink and nextLink parameters are only present if there exists a previous or next page. Check these links to see if more pages of data can be retrieved from the Core Reporting API.

Totals for all results

As mentioned in the pagination information section above, a query to the Core Reporting API can match many rows of data in Google Analytics, but only return a subset of data. The total metric values for all the matched rows are returned in the totalsForAllResults object. This data is useful for calculating averages.

Java

private void printTotalsForAllResults(GaData gaData) {
  Map totalsMap = gaData.getTotalsForAllResults();

  for (Map.Entry entry : totalsMap.entrySet()) {
    System.out.println(entry.getKey() + " : " + entry.getValue());
  }
}

Python

def print_totals_for_all_results(results):
  totals = results.get('totalsForAllResults')

  for metric_name, metric_total in totals.iteritems():
    print 'Metric Name  = %s' % metric_name
    print 'Metric Total = %s' % metric_total

PHP

private function printTotalsForAllResults(&$results) {
  $totals = $results->getTotalsForAllResults();

  foreach ($totals as $metricName => $metricTotal) {
    $html .= "Metric Name  = $metricName\n";
    $html .= "Metric Total = $metricTotal";
  }

  print $html;
}

JavaScript

function printTotalsForAllResults(results) {
  var output = [];

  var totals = results.totalsForAllResults;
  for (metricName in totals) {
    output.push(
        'Metric Name  = ', metricName, '\n',
        'Metric Total = ', totals[metricName], '\n');
  }

  alert(output.join(''));
}

Working Samples

To see the full working samples check out the Core Reporting API Sample in each client library's sample directory.

Java

Google API Java client library Core Reporting API Sample

Python

Google API Python client library Core Reporting API Sample

PHP

Google API PHP client library Core Reporting API Sample

JavaScript

Google API JavaScript client library Core Reporting API Sample

JavaScript Source