ARTIFICIAL INTELLIGENCE (CODE 417)
ACADEMIC SESSION 2025-26
Name: Rishit Sawe
Class: 10
Section: A
Roll Number: 30
Name of the teacher: Monisha Chambail
TOPIC
RULE BASED
PROJECT ON
NASA EXOPLANET
INDEX
[Link] CONTENTS PAGE NO.
1. CERTIFICATE 4
2. SYSTEM REQUIREMENT 5
3. INTRODUCTION 6-7
4. AI CYCLE 8
5. DATA ACQUSITION 9-12
6. DATA EXPLORATION 13-17
7. MODELLING 18
Department of Computer Science
CERTIFICATE
This is to certify that Rishit Sawe, a student of class X has successfully
completed the investigatory project on the topic “NASA EXOPLANET”
during the year 2024-25 in partial fulfillment of the Artificial
Intelligence Practical Examination conducted by CBSE , New Delhi.
Teacher-in-charge Head of the Institution
SYSTEM REQUIRMENT
HARDWARE :
MAIN MEMORY: 500 GB
MICRO PROCESSOR: Intel(R) Core(Tm) i5-8250U CPU @
1.60GHz ,1800 Mhz , 4 Core(s)
SYSTEM TYPE: Latitude 3490
HARD DISK DRIVE: Win_10
CACHE MEMORY: 8 GB
SOFTWARE:
OPERATING SYSTEM : Windows 10
INTRODUCTION
The universe is big — billions of galaxies, every one of them full of stars and, perhaps,
planetary systems of their own. For centuries, humanity has questioned whether the Earth is
unique or if other planets exist to harbor life. This led to an extraordinary domain of space
science: explanatory research.
Exoplanets, or extrasolar planets, are planets that orbit stars beyond our solar system.
Instead of the eight planets orbiting the Sun, exoplanets orbit distant stars light-years away
from us. Although the concept of other worlds has been around since ancient times, the first
exoplanet to be confirmed didn't happen until the 1990s. Since then, the discipline has grown
exponentially. To date, as of 2025, more than 5,600 exoplanets have been confirmed, and
thousands are still pending confirmation.
These exoplanets are highly varied. Some are gas giants much more massive than Jupiter.
Others are rocky, like Earth, or icy, like Neptune. Scientists have even found rogue planets —
planets that wander through space without being orbited by a star. Of all these, the most
interesting are Earth-like exoplanets that orbit in the habitable zone of their stars — the
region where temperatures might be just right for liquid water to exist.
The detection of exoplanets has been facilitated by impressive scientific missions and
sophisticated space telescopes. NASA's Kepler Space Telescope, launched in 2009, was one of
the best planet-finding missions in history. Through the observations of over 150,000 stars,
Kepler detected thousands of exoplanet candidates. Its main technique was the transit
method, which identifies small dips in a star's brightness as a planet passes across it. Even a
faint dimming, when it is done in a regular pattern, may indicate the presence of an orbiting
[Link].
In Kepler's wake, the Transiting Exoplanet Survey Satellite (TESS) launched in 2018 to survey
nearly the entire sky. TESS is still discovering new planetary systems and delivering data that
allow scientists to target planets for further examination. Now, the James Webb Space
Telescope (JWST) is pushing this inquiry still further by examining exoplanet atmospheres for
chemical signatures like water vapor, carbon dioxide, methane, and other possible
indications of life. These findings are enabling scientists not only to wonder where the
planets are, but what they are like. Exoplanet science is not the domain of professional
astronomers alone. Perhaps the most thrilling thing about contemporary science is that it
involves citizens. NASA's Exoplanet Watch and Planet Hunters initiatives enable students and
the general public to engage in actual scientific observation. With free access to data via
services like the NASA Exoplanet Archive, students can download data and employ software
tools to observe for themselves.
In this project, we have used Artificial Intelligence (AI) principles to study NASA's exoplanet
data set. In programming languages such as Python, we have used a rule-based model to
study parameters such as mass, radius, speed of orbit, and method of discovery. The rules
assisted in classifying the planets into types such as Super Earth, Gas Giant, or Neptune-like,
and assisted in determining whether they can be in the habitable zone.
We also made visualizations like bar charts between planet masses and scatter plots of types
of planets in solar systems. Visualizations make large datasets easy to interpret and science
more interesting, particularly for young scientists.
Exoplanet exploration integrates several fields — astronomy, physics, data science, coding,
and design. It does not only inform students about space, but also how to collect evidence,
create hypotheses, interpret data, and make rational conclusions. It promotes collaboration,
imagination, and problem-solving.
In summary, exoplanet study is much more than a scientific fad — it's a frontier. It provides a
peek into the composition of other solar systems and takes us closer to solving the largest
questions of them all: Are we alone? Is Earth special? What else is out there? And as we press
further into the universe, we learn more about our own world — and ourselves.
AI CYCLE
PROBLEM SCOPING 4W CANVAS
WHO – Our Stakeholders:
Our team of Researchers, Scientists, and Astronomers, who are deeply involved in the
exploration of exoplanets, face a significant challenge in today’s era of space exploration. With
the rise of powerful missions like Kepler, TESS, and the James Webb Space Telescope, vast
amounts of data have been collected. However, this data is often fragmented, raw, and lacks a
unified structure that would allow for easy interpretation and meaningful conclusions.
WHAT – The Problem They Face:
They have a problem that, despite the availability of rich exoplanetary data, it is difficult to make
quick and accurate assessments about these distant worlds. The datasets come with missing
values, inconsistent measurement units, varied discovery methods, and lack of visual
representation. This makes it hard to determine important planetary characteristics such as mass,
radius, orbit, or potential habitability — slowing down research progress and scientific
discovery.
WHERE – The Situation or Context:
This problem arises when or while researchers are analyzing exoplanet data from a variety of
sources such as NASA's Exoplanet Archive. The process of cleaning, filtering, comparing, and
interpreting these large datasets manually is time-consuming and prone to error. Without a smart
and automated system in place, researchers face delays in uncovering valuable insights that could
advance our understanding of the universe.
WHY – The Need for a Solution:
An ideal solution would be the development of an AI-powered, rule-based model supported by
data cleaning tools, statistical methods, and graphical visualization techniques. This model would
automate the analysis process — converting raw data into organized information and providing
clear visuals such as bar graphs, scatter plots, and classification summaries. Such a system would
make it easier to identify habitable planets, compare planetary systems, and even engage students
in hands-on space science projects using real data.
DATA ACQUISITION
from selenium import webdriver
from [Link] import By
from bs4 import BeautifulSoup
import requests
import time
import pandas as pd
# NASA Exoplanet Catalog URL
START_URL = "[Link]
catalog/"
# Initialize WebDriver (ensure the WebDriver
executable is in your PATH)
browser = [Link]() # or [Link]()
if using Edge
[Link](START_URL)
# Allow the page to load
[Link](10)
# List to store scraped data
planet_data = []
'''
# Function to scrape data from individual planet
pages
def scrape_planet_details(hyperlink):
try:
page = [Link](hyperlink)
soup = BeautifulSoup([Link],
"[Link]")
temp_list = []
# Extract desired data fields
for tr_tag in soup.find_all("tr", attrs={"class":
"fact_row"}):
td_tags = tr_tag.find_all("td")
for td_tag in td_tags:
try:
temp_list.append(td_tag.find_all("div",
attrs={"class": "value"})[0].contents[0])
except:
temp_list.append("")
planet_data.append(temp_list)
except:
[Link](1)
scrape_planet_details(hyperlink)
# Parse the main page content
soup = BeautifulSoup(browser.page_source,
"[Link]")
# Find all planet entries
planet_entries = soup.find_all("li", class_="exoplanet")
# List to store hyperlinks
planet_hyperlinks = []
# Extract hyperlinks
for planet in planet_entries:
link_tag = [Link]("a")
if link_tag:
planet_hyperlinks.append("[Link]
v" + link_tag.get("href"))
for index, link in enumerate(planet_hyperlinks):
print(f"Scraping data for planet {index + 1}:
{link}")
scrape_planet_details(link)
print(f"Data scraping for planet {index + 1}
completed")
# Define headers based on the data extracted
headers = ["Planet Type", "Discovery Date", "Mass",
"Planet Radius", "Orbital Radius", "Orbital Period",
"Eccentricity", "Detection Method"]
# Create a DataFrame
planet_df = [Link](planet_data,
columns=headers)
# Save to CSV
planet_df.to_csv("exoplanet_data.csv", index=False)
[Link]()
'''
DATA EXPLORATION
Web scrapping
“Code
from selenium import web driver
from [Link] import By
from bs4 import Beautiful Soup
import requests
import time
import pandas as pd
# NASA Exoplanet URL
START_URL = "[Link]
# Webdriver
browser = webdriver. Edge ()
browser. Get (START_URL)
time. Sleep (10)
new_planets_data = []
def scrape_more_data(hyperlink):
try:
page = requests. Get(hyperlink)
soup = Beautiful Soup (page. Content, "html. parser")
temp_list = []
for tr_tag in soup.find_all("tr", attrs={"class": "fact_row"}):
td_tags = tr_tag.find_all("td")
for td_tag in td_tags:
try:
temp_list.append(td_tag.find_all("div", attrs={"class": "value"})[0].contents[0])
except:
temp_list.append("")
new_planets_data.append(temp_list)
except:
[Link](1)
scrape_more_data(hyperlink)
planet_df_1 = pd.read_csv("updated_scraped_data.csv")
for index, row in planet_df_1.iterrows():
print(row['hyperlink'])
scrape_more_data(row['hyperlink'])
print(f"Data Scraping at hyperlink {index+1} completed")
scrapped_data = []
for row in new_planets_data:
replaced = []
for el in row:
el = [Link]("\n", "")
[Link](el)
scrapped_data.append(replaced)
print(scrapped_data)
headers = ["planet_type","discovery_date", "mass", "planet_radius",
"orbital_radius", "orbital_period", "eccentricity", "detection_method"]
new_planet_df_1 = [Link](scrapped_data,columns = headers)
new_planet_df_1.to_csv('new_scraped_data.csv',index=True,
index_label="id")
Data Visualization / Exploration
SOLAR
LIGHT YEARS PLANET PLANET
ID NAME SYSTEM
FROM EARTH MASS TYPE
NAME
3665 Kepler-903 b 2704 4.7 Earths Super Earth KOI-351
3824 KOI-142 b 1106 9.5 Earths Neptune-like KOI-351
3825 KOI-142 c 1106 0.674 Jupiter’s Gas Giant KOI-351
3826 KOI-1599.01 3813 4.6 Earths Super Earth KOI-351
3827 KOI-1599.02 3813 9 Earths Super Earth KOI-351
3828 KOI-1783.01 2896 0.22339 Jupiter’s Gas Giant KOI-351
3829 KOI-1783.02 2896 15 Earths Neptune-like KOI-351
3830 KOI-217 b 2610 Unknown Gas Giant KOI-351
Bar representation:
Plot chart:
Modeling
The model used in this project is based on a rule-based approach, which applies predefined rules to
analyze and classify exoplanet data from NASA’s catalog.
Rule-Based Approach Overview
In a rule-based system, decisions are made according to a set of explicit, logical rules. These rules are
built from domain knowledge and predefined scientific criteria, such as ranges of mass or orbital
distance. This method is both interpretable and deterministic, which makes it ideal for projects where
decisions must be clearly explained and justified.
Application in This Project
For the NASA exoplanet dataset, the rule-based model is used to evaluate each planet using the
following steps:
1. Evaluate Features:
- Uses predefined rules to calculate gravity using the planet’s mass and radius.
- Checks if the planet lies within the 'Goldilocks Zone' (habitable zone) based on its orbital radius and
the temperature of the host star.
- Analyzes orbital speed, eccentricity, and distance from Earth.
2. Feature Extraction:
- Applies logical conditions to determine if a planet qualifies as a Gas Giant, Super Earth, Neptune-like,
or potentially habitable.
- Flags missing or unknown data to ensure consistency.
3. Classification:
- Adds derived features to the dataset, allowing for structured analysis and easy grouping.
- Provides simplified summaries for students and researchers to interpret and explore planetary
systems.
This approach ensures that each planet is evaluated consistently using the same logic and scientific
basis, making the results easy to understand and replicate. It enables structured comparison and helps
identify interesting exoplanets for further study.