This page describes the fundamental knowledge that you need in order to use the Product Sentiment Insights API.
HTTP service and data sets
The Product Sentiment Insights API differs from a traditional "REST" style API in that it consists of two distinct components which are described below:
- HTTP service
The HTTP service refers to the HTTP interface that is used to request product sentiment insights data sets. Interaction with the HTTP service is a multi-step process that begins with requesting a resource that lists available manifest files, then requesting the desired manifest file, and finally requesting the data files as identified in the manifest file. The HTTP service is described in greater detail in the Workflow Walk-Through tutorial.
- data sets
The data sets consist of data files in a compressed format containing product sentiment insights. Data sets are requested using the HTTP service. Unlike a traditional "REST" style API, these data sets are pre-calculated on a regular basis. Further, instead of representing a resource, each data file contains many lines of text that can be evaluated as a JSON object. Learn more about the data sets on the Reference page.
This hybrid architecture allows for speed at the cost of some flexibility. For example, we encourage you to perform multiple concurrent requests to the HTTP service. However, because the data sets are pre-computed, the HTTP service does not offer any filters or sorting options common to "REST" style APIs.
The Product Sentiment Insights API supports the following environments:
||Used while developing your application.|
||Used when your application is complete.|
These environments use different API keys and as a result have independent rate limits. This allows you to develop applications using the staging environment without jeopardizing the stability of the production environment. However, both environments will return production data.
The Product Sentiment Insights API exposes two version types: One for the HTTP service and another for the data sets. This allows you to retain control over how you interact with the Product Sentiment Insights API and decouples your implementation from future changes to the Product Sentiment Insights API until you are ready to upgrade.
This table shows all versions:
|Version||Release Date||Deprecation Date||Status|
HTTP service version (identifies the structure of the URL path, query string parameters and headers)
data sets version (identifies the structure of the data returned by the Product Sentiment Insights API)
- This is newest version. Feature development is focused on future versions, but we may perform backwards compatible bug fixes. New applications SHOULD use this version.
If you are an existing PSI customer, minor modifications may be suitable for your applications if you want to introduce special handling of French/German or incentivized content. You also have the option of requesting the exclusion of incentivized content, and selecting which of the supported languages are included in your PSI datasets.
On July 31 2022, the v1 datasets were augmented to start including Spanish content. The dataset version was not incremented as this is a non-breaking change, and Client PSI Applications are expected to continue working without modifications.
The Product Sentiment Insights API supports the following configuration options:
- Incentivized Content
You can choose whether incentivized content is included in your datasets. Incentivized content will be included by default.
You can choose which of the supported languages are included in your datasets. The supported languages are English, Spanish, French and German. All supported languages will be included by default.
To update the values for these configuration options, please reach out to Bazaarvoice Client Care or your technical contact at Bazaarvoice.
Data update intervals
Product Sentiment Insights API data is available in the following intervals:
|At least monthly||
||Bazaarvoice will generate data sets at least once a month for all available content. You should completely replace your data store using the latest data files as they are available. This ensures that your data store does not fall out of sync with the Bazaarvoice network.|
The following required credentials will be provided by Bazaarvoice:
|Passkey||A unique per-environment identifier included with each request. You will receive two passkeys, one for staging and one for production.|
|Shared Secret||Used to encrypt the access signature. Known only to you and Bazaarvoice. You will receive two secrets, one for staging and one for production.|
Request access to the Product Sentiment Insights API.