Rotterdam impressions

From alwiki
Revision as of 22:07, 25 January 2024 by Alnik (talk | contribs) (Created page with "<div style=" font-family: courier new, monospace; font-size: 14px; "> == <span style="color: white; font-family: Menlo; text-decoration:none; background-color: #ff0033; padding-top: 0.1vw; padding-bottom: 0.1vw; padding-left: 0.1vw; padding-right: 0.2vw;"> _context </span>== This work is a showcase of a tool, developed as part of Experimental Publishing master students' 16th Special Issue [december 2021] - [https://hub.xpub.nl/soupboat/si16/projects/annotation-compa...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

_context

This work is a showcase of a tool, developed as part of Experimental Publishing master students' 16th Special Issue [december 2021] - the Annotation Compass. It builds composites from aggregated vernacular impressions, rich of their subjectivity and situatedness. It is the outcome of a three months journey questioning the relationship between vernacular languages and natural language processing tools.

_about "rotterdam impressions"

We all arrived here from different places, backgrounds, histories and languages. As a group, we are building our relationships and exploring our personalities. Brought to this new place, we discover it both individually and collectively. The city is our everyday environment - we navigate in it to go to school, to go home, to do our grocery, to have a drink out. We spend some of these moments alone, sometimes with someone else. In a short moment of reflection, we all remembered our first or strongest impressions of some particular areas of Rotterdam, which triggered a memorable emotion in us.

These illustrated subjective maps represent the impressions of the first year master students of Experimental Publishing at the Piet Zwart Institute. Their situated experiences trigger various ways to describe their new home: Rotterdam. They were invited to reflect on the way they remember some of the areas in the city and to describe how they feel about them. In words, in their own vernacular English - the common language that is not native to anyone of them but the main language they communicate between each other. Afterwards, their descriptions were filtered by a Python function to extract the most common words they used whilst describing their perceptions of the selected areas on the map, divided in three main categories - adjectives, nouns and pronouns. The final outcome is this short list of the most common impressions of various places and illustrated maps as an artistic expression of the author.

In the following lists we observe the most common words the annotators used in their descriptions. They were asked to describe how they feel about areas and places they see on the map which trigger some memories and emotions in them. By using the Python filters for extracting the words that are adjectives (the words people used to describe their impressions of Rotterdam and various areas and places in it), nouns (the objects that we see people referring to) and pronouns (they way the approached the reflection) in combination with a filter for the most common words (which shows us which words are used the most and how many times), we can observe what are the feelings and impressions that they share as a new collective. What are the places and emotions that are common for some of them?

('I', 53), ('my', 21), ('we', 19), ('me', 16), ('i', 13), ('place', 9), ('super', 9), ('We', 9), ('happy', 8), ('excited', 8), ('our', 8), ('were', 8), ('beautiful', 7), ('Rotterdam', 7), ('first', 6), ('area', 6), ('calm', 5), ('nervous', 5), ('big', 5), ('fig', 5), ('tree', 5), ('safe', 4), ('home', 4), ('inspired', 4), ('city', 4), ('little', 4), ('room', 4), ('shocked', 4), ('fun', 3), ('frustrated', 3), ('friendly', 3), ('good', 3), ('best', 3),  ('bridge', 3), ('school', 3), ('nice', 3), ('station', 3), ('weird', 3), ('sad', 3), ('bike', 3), ('neighborhood', 3)

_a vernacular map

The visual maps of [rotterdam impressions] depict how many times specific words were mentioned in the descriptions overall - the positions of the words are not related to the specific areas where they were mentioned because they are a collective print of how many times these words were used anywhere on the map. By printing each time a word was used, we can see the intensity of words used by the annotators. They are coloured due to their category - adjectives are in green, nouns are in orange and pronouns - in pink.

how are we describing

"I" and "we" are the most common pronouns (in pink) that were used in the annotations. You can see that "I" was used as twice more often that "we", but perhaps because of the context of us describing our first impressions of a new place where everyone arrived to study and work together as a group, some of the experiences were still connected to a collective moment.

adjectives as emotions and impressions; nouns as objects

In the context of the annotation compass, the adjectives (in green) were filtered and extracted as impressions and descriptions of the new place where people arrived. They are either describing Rotterdam and places inside it, or the way the annotator feels about them. Either way, in most of the cases we can still get the emotion that lays behind the word used. The most common nouns are understood as the objects (in orange) - the ones people are relating most often to.

_functions

to filter the adjectives from the annotations:

def adjectives(text):
   adj = []
   for (word, tag) in labels_blob.tags:
       if tag == 'JJ':
           adj = adj + [word]
   return adj


to filter the nouns from the annotations:

def nouns(text):
   noun = []
   for (word, tag) in labels_blob.tags:
       if tag == 'NN':
           noun = noun + [word]
   return noun

to filter the most common words from the annotations:

def most_common(words):
   freq_dist_pos = FreqDist(words)
   return freq_dist_pos.most_common(100)