Analysis Of Forest Fire Predictors In Montesinho Natural Park

Posted on Jul 25, 2016

Montesinho is a beautiful protected area located in the municipalities of Vinhais and Bragança, northeastern Portugal. Sections of the southern slopes of the Serra da Coroa (Sierra de la Culebra) fall within the park.


It is home to many different kinds of animals. Its biodiversity includes the Iberian wolf, roe deer, wild boar, Iberian lynx, common genet, red fox and European otter.




What a disaster it would be if there were a forest fire!

Today I am going to analyze the Forest Fire Predictors In Montesinho Natural Park. The forest fire data concerns burned areas of the forests in Montesinho Natural park due to forest fires. It was collected from January 2000 to December 2003 . It contains 517 instances

13 variables (1 dependent variable, 4 discrete attributes and 8 continuous attributes). These are the variables:

  • Response: area: Burned area of a forest fire (ha)
  • Predictors:
  • X: x-axis coordinate of the Montesinho park map: 1 to 9
  • Y: u-axis coordinate of the Montesinho park map: 2 to 9


    • FFMC: A numerical rating of the moisture content of litter and other cured fine fuels: 18.7 to 96.2
    • DMC: A numerical rating of the average moisture content of loosely compacted organic layers and medium-size woody material: 1.1 to 291.3
    • DC: A numerical rating of the average moisture content of deep, compact, organic layers: 7.9 to 860.6
    • ISI: A numerical rating of the expected rate of fire spread: 0.0 to 56.10
    • Month: month of the year: 1 to 12
    • Day: day of the week: 1 to 7
    • Temp: temperature in Celsius degrees: 2.2 to 33.30
    • RH: relative humidity in %: 15.0 to 100
    • Wind: wind speed in km/h: 0.40 to 9.40
    • Rain: outside rain in mm/m2: 0.0 to 6.4

    Here we raise some questions:

    • How are the forest fires distributed in the park?
    • What are the significant variables to predict forest fires?
    • How are these variables related to the area of forest fires?
    • What can we suggest to tourists and the fire department?

    First, how are forest fires distributed in the park.


How large can a forest fire be? We did a summary on the burned down area data and categorized the area as 'small' if the area is under first quantile, as 'median' if the area is between first and third quantile, as 'large' if it if above the third quantile. So the summary is the following:


Min. 1st Qu.  Median Mean 3rd Qu. Max.

0.09 2.14 6.37   24.60   15.42 1091.00

Categorize area by the above information:

(0,2.14)  size of area = ‘small’

(2.14,15.42)  size of area = ‘median’

(15.42,1091.00) size of area = ‘large’

We did a boxplot of each category and found out that there are some outliers. After taking out the outliers, we zoom in and do a box plot for category median and small.

Rplot04         Rplot05

Second, we come to the question: which variables are the most significant? In order to achieve this goal, I did some statistical analysis. It turns out that the

fit = glm(log(area+1)~FFMC+DMC+DC+ISI+temp+RH+wind+rain,data=mydata_new,family=Gamma())

> summary(fit)


glm(formula = log(area + 1) ~ FFMC + DMC + DC + ISI + temp +

RH + wind + rain, family = Gamma(), data = mydata_new)

Deviance Residuals:

Min 1Q Median 3Q Max

-2.11417 -0.52704 -0.08414 0.28686 1.39097


Estimate Std. Error t value Pr(>|t|)

(Intercept) 5.405e-01 6.776e-01 0.798 0.426

FFMC -2.129e-03 7.666e-03 -0.278 0.781

DMC -6.427e-04 3.960e-04 -1.623 0.106

DC 8.486e-05 1.036e-04 0.819 0.413

ISI 1.043e-02 6.706e-03 1.556 0.121

temp 7.194e-04 4.458e-03 0.161 0.872

RH 1.994e-03 1.420e-03 1.405 0.161

wind -1.043e-02 1.004e-02 -1.039 0.300

rain -1.099e-02 4.172e-02 -0.263 0.792

Based on the above analysis, DMC, ISI,RH are the three most significant variables, having critical values of 0.106,0.121 and 0.161 respectively.


Rplot08 Rplot07


Rplot10 Rplot09


Rplot12 Rplot11

What can we suggest to tourists and to the fire department? Summer and Fall are the seasons when there are the most tourists, thus we suggest tourists be more careful when using fire (camping, BBQ etc).The fire department should look closer at measurements of RH, ISI, DMC variables and prepare accordingly.

One of the drawbacks of this data set is that it does not record human activity which, I believe, does have a huge impact on the occurrence of forest fires.

What times do forest fires happen the most?


Indeed, fall and summer are the most dangerous times due to high temperature and low relative humidity. Most importantly, people usually go camping and hiking in this two seasons, which brings potential danger.

Next steps:

  1. More closely analyze the relationship of variables to each other.
  2. Develop  a more precise model for predicting the area based on the variables we have now.

I will keep on improving this project. Thank you very much for reading this post. Please feel free to give any advice!


About Author

Miaozhi Yu

Miaozhi recently received her Master’s degree in Mathematics from New York University. Before that she received a Bachelor’s Degree in both Mathematics and Statistics with a minor in Physics from UIUC. Her research interests lie in random graphs...
View all posts by Miaozhi Yu >

Leave a Comment

No comments found.

View Posts by Categories

Our Recent Popular Posts

View Posts by Tags

#python #trainwithnycdsa 2019 airbnb Alex Baransky alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus API Application artist aws beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep Bundles California Cancer Research capstone Career Career Day citibike clustering Coding Course Demo Course Report D3.js data Data Analyst data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization Deep Learning Demo Day Discount dplyr employer networking feature engineering Finance Financial Data Science Flask gbm Get Hired ggplot2 googleVis Hadoop higgs boson Hiring hiring partner events Hiring Partners Industry Experts Instructor Blog Instructor Interview Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter lasso regression Lead Data Scienctist Lead Data Scientist leaflet linear regression Logistic Regression machine learning Maps matplotlib Medical Research Meet the team meetup Networking neural network Neural networks New Courses nlp NYC NYC Data Science nyc data science academy NYC Open Data NYCDSA NYCDSA Alumni Online Online Bootcamp Online Training Open Data painter pandas Part-time Portfolio Development prediction Prework Programming PwC python Python Data Analysis python machine learning python scrapy python web scraping python webscraping Python Workshop R R Data Analysis R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn Selenium sentiment analysis Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau team TensorFlow Testimonial tf-idf Top Data Science Bootcamp twitter visualization web scraping Weekend Course What to expect word cloud word2vec XGBoost yelp