Uncovering Hidden Players in the Space Economy
Researchers identify new companies in the growing space sector through innovative methods.
Kenza Bousedra, Pierre Pelletier
― 7 min read
Table of Contents
- A New Method to Find Companies
- The Approach
- Gathering Information from the Press
- Why Newspapers?
- Collecting Articles
- The Rules for Identifying Companies
- Creating the Rules
- Putting the Method to the Test
- Initial Findings
- New Companies Discovered
- By-Products of the Method
- Evaluation of the Method
- How Well Did It Work?
- Comparison with Statistical Methods
- Future Directions for Improvement
- Room for Growth
- Exploring Co-citation
- Conclusion
- Final Thoughts
- Original Source
- Reference Links
The space economy is becoming increasingly important as new technologies and services emerge. More and more companies are getting involved in what is called "downstream space activities." This refers to businesses that use data and technology from space, like satellites, to create services or products. Examples include satellite-based internet, Earth monitoring, and navigation services.
But here's the catch: many of these companies are not well-known or are not easily classified under traditional industry categories. This makes it hard for experts to keep track of them. So, researchers decided it was time to find a way to identify these companies more efficiently.
A New Method to Find Companies
To tackle this issue, a group of researchers devised a method to spot companies involved in downstream space activities. The aim was to create a list of companies that might have escaped the radar of traditional identification methods.
The Approach
The researchers used something called Named Entity Recognition (NER). In simple terms, it is a tool that helps computers understand and recognize names within text. Instead of relying on complicated statistical methods, this approach focuses on rules set by experts in the field.
The researchers began their work by taking a close look at newspaper articles. Newspapers are rich sources of information, and they often mention companies involved in space activities. By gathering a large number of articles about space, they hoped to find names of new companies that are involved in the sector.
Gathering Information from the Press
Why Newspapers?
Why did the researchers choose newspapers? Well, let's face it—if you want to know what’s happening in the world, you turn to the news! Newspapers cover a wide variety of topics, including business and technology, and they update quickly with the latest information.
The researchers decided to use both French and English articles. They focused on a time period starting from 2000 to the present day. By covering two decades, they increased their chances of finding companies that may have been overlooked before.
Collecting Articles
The researchers gathered articles from two major databases, Factiva and Europresse. These platforms compile news articles from different sources, making it easier to find relevant content. Using specific search queries, they retrieved thousands of articles that mentioned space activities.
In total, they managed to collect about 49,000 articles. After sorting through them, they focused on the ones that specifically talked about downstream activities in space.
The Rules for Identifying Companies
Creating the Rules
Now that they had a mountain of articles, the researchers needed a smart way to sift through this information. They crafted a series of rules to help identify potential companies involved in downstream space activities.
Industry Codes and Legal Categories
Rule 1:The first rule was about filtering companies based on their industry codes and legal status. The researchers created a “dictionary” that only included companies that belonged to specific legal categories related to space activities.
For example, companies involved in satellite communications or Earth observation were included, while unrelated businesses were filtered out. This made the identification process much more efficient.
Rule 2: Capital Letters Only
It's common for company names to start with capital letters. So, the researchers decided to keep only words that started with capital letters in the text. This move helped reduce the clutter from regular words and phrases, making it easier to spot potential company names.
Rule 3: Context Matters
To avoid picking up random names that didn’t actually refer to companies, the researchers added a third rule. They looked for names that appeared within a specific context—basically, a certain distance from common words or phrases associated with space activities. This helped to further narrow down the list of candidates.
Rule 4: Regular Expressions
Finally, the researchers employed a rule related to "regular expressions." Essentially, they looked for patterns in company names that often appeared in the space sector. By identifying these patterns, they could filter out names not likely to belong to companies in the downstream space market.
Putting the Method to the Test
With their rules in place, the researchers applied the method to the collected articles. They were eager to see if they could identify new companies that hadn’t been documented before.
Initial Findings
After applying their rules, the researchers discovered 1,475 potential companies. From this list, they conducted a manual review to confirm which ones were indeed involved in downstream space activities.
Generally, this final check involved revisiting the articles where these companies were mentioned or checking their websites to see what services they offered.
New Companies Discovered
As a result of their method, the researchers identified 88 new companies in the downstream space sector. They were thrilled with the outcome, as it showed that the method could successfully spot previously unknown players in the industry.
By-Products of the Method
Interestingly, they also discovered 30 additional companies that weren't on their list but came up during the review process. These companies were found because they were mentioned alongside the identified firms.
Essentially, the researchers realized that knowing one company often led to discovering another, like finding hidden gems while sifting through a treasure chest!
Evaluation of the Method
How Well Did It Work?
After identifying new companies, the researchers evaluated their method. They looked at how many known companies they could find and whether their rules effectively filtered out irrelevant names.
They managed to preserve a significant number of known companies while minimizing noise—a win-win for their approach!
Comparison with Statistical Methods
To see how their method stacked up against other techniques, they compared it with a statistical NER model. Surprisingly, even though the statistical model flagged a lot of companies, it failed to match as many known companies compared to the rule-based method devised by the researchers.
In short, the researchers found that their custom rules were more efficient than the generic statistical model. So rather than just throwing spaghetti at the wall and seeing what sticks, they created a precise set of criteria that produced better results.
Future Directions for Improvement
Room for Growth
While the researchers were pleased with their results, they acknowledged that there was still room for improvement. For instance, they noted that their method relied heavily on human expertise. To increase efficiency, they suggested possibly incorporating machine learning techniques that could learn from new data over time.
Co-citation
ExploringThe researchers also highlighted the interesting phenomenon of co-citation—that is, companies being mentioned together in the same articles. They thought it would be worth investigating these relationships further, as understanding connections between companies could provide valuable insights into the space sector.
Conclusion
In summary, identifying new companies in the downstream space sector is no small feat. The researchers developed a practical, rule-based method to uncover hidden players in this ever-growing industry. With their approach, they managed to spot 88 new companies, enriching the existing database and paving the way for future exploration.
And who knows? The next time you're checking your satellite-enabled GPS, you might just be utilizing the innovation brought forth by one of the newly discovered companies!
Final Thoughts
While the journey to identify new companies in the space sector is ongoing, the researchers' approach showcases the importance of adapting to new challenges in the business world. By combining expertise and cutting-edge techniques, they set a precedent for future efforts in understanding the evolving landscape of the space economy.
With every new company identified, the skies are not the limit—they're just the beginning!
Original Source
Title: A Rule-Based Methodology for Company Identification: Application to the Downstream Space Sector
Abstract: This paper proposes an original methodology based on Named Entity Recognition (NER) to identify companies involved in downstream space activities, i.e., companies that provide services or products exploiting data and technology from space. Using a rule-based approach, the method leverages a corpus of texts from digitized French press articles to extract company names related to the downstream space segment. This approach allowed the detection of 88 new downstream space companies, enriching the existing database of the sector by 33\%. The paper details the identification process and provides guidelines for future replications, applying the method to other geographic areas, or adapting it to other industries where new entrants are challenging to identify using traditional activity classifications.
Authors: Kenza Bousedra, Pierre Pelletier
Last Update: 2024-12-03 00:00:00
Language: English
Source URL: https://arxiv.org/abs/2412.02342
Source PDF: https://arxiv.org/pdf/2412.02342
Licence: https://creativecommons.org/licenses/by/4.0/
Changes: This summary was created with assistance from AI and may have inaccuracies. For accurate information, please refer to the original source documents linked here.
Thank you to arxiv for use of its open access interoperability.
Reference Links
- https://www.data.gouv.fr/fr/datasets/base-sirene-des-entreprises-et-de-leurs-etablissements-siren-siret/
- https://www.entreprises.gouv.fr/fr/actualites/crise-sanitaire/france-relance/france-relance-premiers-laureats-du-volet-spatial
- https://www.connectbycnes.fr/en/space-for-good
- https://commercialisation.esa.int/startups/
- https://spacy.io/universe/project/video-spacys-ner-model