Mining and Analyzing Facebook Data

Execute the plan of attack for Facebook data analysis by extracting basic page data, obtaining access tokens, and graphing followers, engagement, likes and dislikes, and demographics to guide marketing.

Create a Facebook app and generate a Graph API access token to start mining and analyzing Facebook data. Learn to configure app basics, roles, and permissions for data access.

Use the Facebook Graph API with Python to fetch data from your page, selecting fields and posts messages. Understand permissions and privacy as you access page descriptions and post content.

Learn how to obtain a temporary one-hour Facebook access token in Python via Google Colab, install the Facebook SDK, configure app id and secrets, and exchange for a long-lived token.

Learn how to obtain a permanent page access token for a Facebook page using the Graph API, including token, page id, and API version setup.

Extract basic profile data from a Facebook page using the Graph API, with a permanence based token and page ID, retrieving fields such as name, category, about, and website.

Fetch Facebook page fans by date using the Graph API and Page Insights for the last 90 days, then visualize with a pandas data frame and a line graph.

Explore fans by language in Facebook data mining, visualizing daily page fans by language and creating a data frame to compare languages like Portuguese, English, and Spanish.

Analyze Facebook fans by city using a data frame, plot graphs and a map to reveal cities with the most fans, highlighting Brazilian cities like Belo Horizonte and Brasília.

Visualize fans by country using aggregated counts, render a bar chart and maps, and rename the data frame to show country and total fans.

Explore fans by age and gender on a Facebook page, extract the data into a data frame, and visualize it with bar plots to reveal the leading demographics.

Analyze daily likes and dislikes for a Facebook page, compute per-day values, and visualize the data to compare trends.

Analyze how likes and dislikes relate on a Facebook page by counting actions and identifying reasons behind dislikes, including negative actions, and visualize the results with a bar graph.

Analyze Facebook data metrics, visualizations, engagements, and impressions while exploring clicks on contact information, and apply additive regression to forecast future page performance.

Visualize Facebook page tab views by comparing daily and weekly results. Build data frames and bar charts to highlight which tabs attract the most engagement over time.

Analyze the clicks on contact information and call-to-action links on Facebook, across daily, weekly, and 28-day periods. Create and visualize a clicks_contacts metric to guide marketing strategy.

Explore how Facebook page users click the directions button to find your business, and analyze metrics like page gets directions, logged in unique counts, and calls to optimize engagement.

Explore the page engagement metric by examining page engaged users and total engagements from daily to 28 days, and visualize data for future machine learning.

Explore engagement by type on Facebook data, analyzing consumption type metrics such as link clicks, video plays, and photo views across daily, weekly, and 28-day intervals.

Discover how the page check-in metric counts visits to your physical location, with views by day or week, and how check-ins appear on users' timelines.

Explore positive actions by type on a Facebook page, aggregating by day or week to reveal which actions—comments, likes, shares, or link clicks—drive engagement.

Explore negative actions by type using the metric page, focusing on hide post, hide all posts, and report as spam across 28 days, with results and date aggregation for visualization.

Learn to read visualizations by hour showing how many Facebook fans saw posts, identify peak online times, and plan post schedules for higher visibility.

Analyze paid and unpaid likes by comparing bait and no-bait engagement, visualize page likes over time, and explore related metrics in a data frame.

Explore how page impressions distribute across paid, organic, and viral channels, compare them with normalization and visuals, and preview time series and machine learning insights.

Learn to work with time series data by installing libraries, preparing a data frame with dates and metrics, and using arima to predict future fans and engagement.

Explore time series decomposition of Facebook data, extracting trends, seasonal patterns, and residuals, visualize components, and prepare for future predictions using cleaned datasets.

Apply an arima model to facebook page data to forecast the next 30 days, optimize parameters, and produce a date-indexed prediction of fan counts.

Evaluate time series predictions for Facebook page fans by comparing real data with forecasted values, computing mean absolute error in Python to assess model performance.

Visualize the predictions with a time-series graph, plotting the next 30 days against real data to compare algorithm performance and discuss data from 2020–2021.

Explore time series forecasting with Facebook Prophet to predict engagement on a Facebook data frame from 2021–2022, including daily seasonality and 30-day predictions.

Evaluate Facebook Prophet time series predictions against real engagement data, visualizing daily results and computing mean absolute error to compare Prophet with ARIMA and Facebook for fans.

Plan of attack guides analysts to extract reactions, video views, and post texts from Facebook pages, apply word clouds, keyword searches, and sentiment analysis for marketing insights.

Analyze Facebook post reactions by querying the graph api, counting like, love, wow, anger, and other reactions over daily and weekly intervals to inform posting strategy.

Analyze video playback on your Facebook page by comparing three-second and thirty-second views to measure ad engagement and audience retention.

Explore stories by type in Facebook data, using metrics such as stories by page story type to analyze fan, page post, and other categories over 28 days.

Extract texts from Facebook page posts using graph connections, capturing messages, created times, and links, then organize results in a data frame and analyze post likes.

Install and import the Google translator library to translate texts from Brazilian Portuguese to English. Create a new translation field in the data frame and translate each post.

Develop a text preprocessing function to clean English texts for NLP, including lowercasing, link removal, and stop-word punctuation filtering, and apply it to dataset to create a preprocessed translation column.

Generate a word cloud from pre-processed Facebook post texts to visualize the most frequent words, using a word cloud library, and showing topics like computer vision and AI.

Learn to search posts using spaCy by building tokens and search strings, applying the phrase matcher to locate words and their surrounding context in Facebook data.

Apply a Python workflow to search Facebook posts for keywords like deep learning and Python, using preprocessing and translation attributes to identify and visualize matches in Google Colab.

Aggregate Facebook posts by date to count posts per month and visualize trends. Group by created time, convert to datetime, and plot posts by month to compare January through March.

Extract comments from Facebook posts using graph connections, build a data frame of commands from messages, and prepare for preprocessing and translation in the next lecture.

Learn to pre-process Facebook comments by translating texts to English, cleaning non-text entries, and filtering by length to prepare data for sentiment analysis.

Analyze user comments with sentiment analysis using the Vader lexicon and natural language processing to classify polarity as positive, negative, or neutral, and visualize results.

Recap techniques for extracting Facebook page data, generating visualizations of cliques, engagements, and impressions, using the Showtime series algorithm to predict future engagement, and analyzing posts with natural language processing.

Explore artificial intelligence and data science through AI Expert Academy's online platform, with monthly new courses, certificates, and high-quality videos on machine learning, deep learning, and more.