Module 1 - DataSift PYLON™ for Facebook Topic Data

PYLON for Facebook topic data enables DataSift customers to derive aggregate insights about audiences, brands and topics from social interactions.

Learning Objectives

After completing this module you will be able to:

• Describe how the PYLON model uses the DataSift platform

• List the interfaces that access PYLON data

Content

• Describe PYLON Model

• Describe Interfaces

• Exercise

Resources

• Documentation

• REST API Endpoints

The PYLON model uses the DataSift platform to filter and classify interactions from a source into a private index. Analysis queries are made against interactions in the index. The result is statistical information formatted in JSON or displayed in charts.

Sources

Sources are the social networks, news sites, forums, comments, message boards, blogs and other networks that provide input interactions for the DataSift platform. Facebook is currently the only source available using the PYLON model.

Augmentations

Augmentations add extra data to interactions by the DataSift platform.

Interaction Filtering

Public and non-public data is filtered from a source into a sequestered, privacy-safeguarded index using keywords, demographic characteristics, location information and other attributes defined by DataSift’s powerful Curated Stream Definition Language (CSDL) filtering language.

Classifying Interactions

DataSift VEDO feature classifies interactions by adding user-defined data to interactions in the form of text strings (tags) before an interaction is added to the index.

Recording

Recording is the process of applying an interaction filter to the stream of interactions, and saving matching interactions to the index. If a recording is left running, the most recent 32 days of filtered interactions are held in the index.

Analysis Query

An Analysis Query is used to query the interactions in the index to generate reports.

Analysis Results

There are two types of analysis available for an Analysis Query:

1. Frequency Distribution

2. Time Series

Frequency Distribution

This type of analysis returns the top values for the requested target. The author age target is used in this example.

Time Series

This type of analysis returns the number of interactions or the number of unique authors generating those interactions, where the volume is represented over a regular time interval, based on a timestamp field.

The DataSift platform provides two methods to access PYLON:

1. Web application

2. Application Programming Interface (API)

Web Application

Access PYLON using the web application:

1. Navigate to https://app.datasift.com and sign in to the DataSift platform.

2. Select PYLON tab to create interaction filters, start recordings and generate anaysis charts.

Application Programming Interface (API)

Access PYLON using the pylon family of REST endpoints. The endpoints provide programmatic access to the DataSift platform.

PYLON REST endpoints are covered in more detail in other modules.

Knowledge Check

• Which is the correct order of tasks? Choose one.

  a. Analyze interactions, Classify interactions, Filter interactions, Deliver interactions to an Index

  b. Classify interactions, Filter interactions, Analyze interactions, Deliver interactions to an Index

  c. Filter interactions, Classify interactions, Deliver interactions to an Index, Analyze interactions

  d. Deliver interactions to an Index, Filter interactions, Analyze interactions, Classify interactions

• What DataSift features add extra data to interactions?

• What is the maximum amount of time filtered interactions are stored in an index?

• Which analysis produces the top values for a requested target?

• Which PYLON REST endpoint displays running or stopped interaction filters?

DataSift PYLON Workshop is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.