title: "IDS Workshop - Quanteda"

subtitle: "How we met Quanteda - Analyzing the TV show ‘How I Met Your Mother’ with quanteda"

author: "Alexander Kraess, Augusto Fonseca, Jorge Roa"

date: "21/11/2022"

output:

html_document:

toc: TRUE

df_print: paged

number_sections: FALSE

highlight: tango

theme: lumen

toc_depth: 3

toc_float: true

self_contained: false

editor_options:

markdown:

wrap: sentence

```{r setup, include=FALSE}

```{r, fig.align='left', echo=F, out.width = "30%"}

```{r, fig.align='center', echo=F, out.width = "60%"}

Quanteda is a brilliant package, and we hope that we can show you this power fullness!

In this session You will practice the basic concepts, some filters and also some visualization features.

Fell free to access our GitHub page (https://github.com/jurjoroa/text_analysis_quanteda) to check our code and other details.

**Step 1** Load the packages which we will use (don't forget to install them before!).

```{r, message=F}

If you don't want to go through this, you might go to the session "3. Alternatively... import the database from our GitHub (https://github.com/jurjoroa/text_analysis_quanteda)!".

**Step 2.1** Scrap the website "Springfield! Springfield!" to get the episodes transcriptions.

Springfield! Springfield! hosts a database containing thousands of TV show episode scripts and movie scripts.

Read more: https://www.springfieldspringfield.co.uk/

```{r, message=F}

**Step 2.2 ** Scrap Wikipedia to get the episodes and characters metadata and clean it!

```{r, message=F}

**Step 2.3** Load TV scripts (txt saved files) and merge data

```{r, message=F}

**Step 2.4** OPTIONAL: Save the final dataframe to be used as a corpus text

```{r, message=F}

If you went through session 2, you already get the data raw and can jump to session "4. Uploading data into Quanteda corpus"

**Step 3.1** Load the files

```{r, message=F}

OK! It's showtime! Let's upload all our data into a Quanteda corpus element.

**Step 4.1** First Step: Define a corpus

```{r, message=F}

**Step 4.2** Convert corpus into tokens and wrangle it

**Step 4.3** Create a DFM.

Let's check the new subsetted DFM

```{r, eval=F}

Now that we already upload the data and created the Quanteda elements, we can try some basics analysis.

**Step 5.1** Have you ever thought that you had seen an episode before?

Let's check the similarity between episodes!

```{r, eval=F}

**Step 5.3** Is there a correlation between the episodes?

```{r, eval=F}

**Step 5.4** What is the main actor for you? Does it depend on the season?

```{r, eval=F}

**Step 5.5** What is the most frequent characters that appears in the TV show?

```{r, eval=F}

**Step 5.6** What about the SECONDARY characters?

```{r, eval=F}

SpaCyr is a package where we can use the amazing functions of spaCy for analysis of text.

**Step 6.1** Explanation

```{r, eval=F}

**Step 6.2** Get frequency of adjectives

```{r, eval=F}

**Step 7.1** Network plot

```{r, eval=F}

**Step 7.2** Network plot with 30 principal characters

```{r, eval=F}

**Step 7.3** Network plot of Ted

```{r, eval=F}

**Step 8.1** Text stat collocation

```{r, eval=F}

**Step 9.1** Locate keywords-in-context - kwic function

```{r, eval=F}