Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/Gema-Villanueva/proyecto-eda-roles-datos/llms.txt

Use this file to discover all available pages before exploring further.

Technology demand in the Spanish data job market can be understood from two complementary angles: what skills employers list in job postings, and what technologies the global developer community is actively using and aspiring to learn. This page covers the latter, derived from the Stack Overflow Developer Survey 2025.

​
Data Source

The Stack Overflow Developer Survey 2025 gathered responses from approximately 90,000 developers worldwide. It is one of the largest and most cited annual surveys of the global software development community. Within this project, the survey data was loaded as df_stack and processed to generate the technology_rankings.csv file.
Stack Overflow survey data reflects a global developer community and is not a direct proxy for Spanish job market demand. It functions as a reference signal for broad technology adoption trends — particularly useful for understanding which AI tools are gaining traction. See the Limitations page for a discussion of Stack Overflow-specific biases.
The dataset covers technologies organised into named categories. Each technology has two record types:
  • used — technologies the respondent currently uses in their work or personal projects.
  • wanted — technologies the respondent does not currently use but wants to work with in the coming year.
This distinction enables a gap analysis: technologies with high wanted but lower used counts signal emerging or aspirational technologies. Technologies with high used but lower wanted counts may indicate established tools that are widely adopted but not exciting to practitioners.

​
Technology Categories

The technology_rankings.csv dataset covers the following six categories:

ai_model_tool

AI and LLM tools: GPT models, Claude, Gemini, reasoning models, and similar AI assistants used by developers.

language

Programming languages: Python, JavaScript, TypeScript, SQL, Rust, Go, and others.

database

Databases and data stores: PostgreSQL, MySQL, MongoDB, Redis, Snowflake, BigQuery, and more.

platform_cloud

Cloud platforms and operating environments: AWS, Azure, GCP, and associated managed services.

web_framework

Web frameworks and libraries: React, Django, FastAPI, Node.js, and similar tools.

development_environment

Development environments and IDEs: VS Code, Neovim, JetBrains IDEs, GitHub Copilot, Claude Code, and similar tools.

​
Top AI Model Tools (Used)

The ai_model_tool category is particularly relevant to the data roles landscape. OpenAI’s GPT models lead by a substantial margin among tools currently used by survey respondents.
RankTechnologyUsed Count
1openAI GPT (chatbot models)13,424
2Anthropic: Claude Sonnet7,063
3Gemini (Flash general purpose models)5,823
4openAI Reasoning models5,716
The gap between rank 1 (openAI GPT at 13,424) and rank 2 (Anthropic: Claude Sonnet at 7,063) is substantial — GPT models have approximately twice the usage count of the next most-used AI tool. This likely reflects both early market entry and broad enterprise integration of the OpenAI API.

​
Used vs. Wanted Gap Analysis

The gap between used and wanted counts reveals which technologies are gaining momentum versus which are plateauing. Block 4 of Notebook 04 (04_tecnologias.png) visualises this gap across all technology categories. Interpreting the gap:
  • Positive gap (wanted > used): Emerging technology. Developers want to adopt it but haven’t yet. Strong signal for future demand growth.
  • Near-zero gap: Stable technology. Adoption and aspiration are balanced. Mature, well-established tools often fall here.
  • Negative gap (used > wanted): Established or declining technology. Widely used but fewer developers are aspiring to work with it going forward.
For career planning and curriculum decisions, technologies with a strongly positive gap are worth prioritising — they represent where the community expects to move next, which often leads job market demand by 12–24 months.

​
Skill and Technology Overlap

The project also generates a skill_technology_overlap_eda.csv file that cross-references skills found in job offer postings with technologies catalogued in the Stack Overflow survey. This overlap dataset enables questions such as:
  • Which Stack Overflow technologies appear most frequently in Spanish job postings?
  • Are there high-demand Stack Overflow technologies that are under-represented in Spanish job descriptions?
  • Which job-posting skills have no clear equivalent in the Stack Overflow taxonomy?
The overlap analysis bridges the gap between community-reported technology adoption (df_stack) and employer-stated hiring requirements (df_jobs, df_tecno, df_scraping).

​
Working with the Rankings Data

The following code demonstrates how to load technology_rankings.csv and perform the used vs. wanted gap analysis programmatically: 
import pandas as pd

df_rankings = pd.read_csv('data/clean/technology_rankings.csv')

# Top used technologies
used = df_rankings[df_rankings['type'] == 'used'].sort_values('count', ascending=False)
print(used.head(10))

# Top wanted technologies
wanted = df_rankings[df_rankings['type'] == 'wanted'].sort_values('count', ascending=False)
print(wanted.head(10))

# Used vs Wanted gap
gap = df_rankings.pivot_table(index='technology', columns='type', values='count', fill_value=0)
gap['gap'] = gap['wanted'] - gap['used']
print(gap.sort_values('gap', ascending=False).head(10))
ColumnTypeDescription
categorystringTechnology category (e.g., ai_model_tool, language, database, platform_cloud, web_framework, development_environment)
typestringEither used or wanted
technologystringName of the technology as reported in the Stack Overflow survey
countintegerNumber of survey respondents who selected this technology for the given type
To focus the analysis on a specific category, filter on the category column before pivoting:
import pandas as pd

df_rankings = pd.read_csv('data/clean/technology_rankings.csv')

# Restrict analysis to AI model tools only
df_ai = df_rankings[df_rankings['category'] == 'ai_model_tool']

gap = df_ai.pivot_table(
    index='technology',
    columns='type',
    values='count',
    fill_value=0
)
gap['gap'] = gap['wanted'] - gap['used']
gap['gap_pct'] = (gap['gap'] / gap['used'].replace(0, 1)) * 100

print(gap.sort_values('gap', ascending=False))

​
Visualisation Reference

Block 4 of Notebook 04 (04_tecnologias.png) renders the full used vs. wanted comparison as a paired bar chart. Technologies are grouped by category and sorted by used count within each group. The visual gap between the used and wanted bars is the primary analytical output of this block. This visualisation should be read alongside the job postings skill rankings (derived from job_skills_long) to form a complete picture of the technology landscape — one reflecting what employers ask for, the other reflecting what the developer community is doing and aspiring to.

Build docs developers (and LLMs) love

Get started for freeTalk to us