Skip to content

portfolio — data analyst · FinOps · UOC · Ironhack

AroaXinping

Data analyst. Fine Arts, Dietetics and personal trainer, then the pivot to data science — Data Science degree at UOC and FinOps bootcamp at Ironhack. The odd path makes sense: I know how to read a number before formatting it and I know when a visualisation lies.

FinOps as my niche because it connects the technical with the economic: data that reduces cost, infrastructure that can be justified with numbers. Not analysis for analysis — analysis that changes a decision.

every day
pythonpython
pandaspandas
jupyterjupyter
gitgit
in projects
sqlsql
numpynumpy
scikit-learnscikit-learn
matplotlibmatplotlib
seabornseaborn
plotlyplotly
streamlitstreamlit
prefectprefect
aws s3aws s3
astroastro
borrowed syntax
rustrust
c++c++
shellshell
IronhackData Analytics Bootcamp · 2026Data analysis, SQL, visualisation, applied machine learning.
UOCBSc in Data Science · 2025 – presentIn progress. Statistics, machine learning, data infrastructure.

Specialist analyst in FinOps. Small team, real problem, infrastructure cost that can be moved. If the analysis changes nothing, it is surplus.

  • leaked-key-impact

    AWS keys leaked on purpose to see who uses them and how fast. 1,301 real attacks, mapped to MITRE ATT&CK, enriched with OSINT and with a cost model on top.

    pythonawsstreamlit
    see project →code →

  • cloud-cost-predictor

    ML system that scores 123,000 real VMs and surfaces $5.9M/month in cloud waste. Quantile regression with asymmetric loss and dual CPU + memory models to avoid terminating what is actually busy.

    pythonxgboostplotly
    see project →code →

  • drive-failure-predictor

    Predictive maintenance over SMART data from 318,000 production drives: it detects 96% of failures before they happen, with a cost-calibrated threshold.

    pythonxgboostpolars
    see project →code →
see all 14 projects →