Econf.ai-A mission driven app to make Earth a better place

A platform that factors in organizers’ preferred locations and the participants’ locations, and recommends the optimal location to host the event, which has the lowest flight costs and emissions.

Comment

Inspiration

Arranging logistics for any event is a challenge. Especially, when you factor in the huge amount of participants flying in from different parts of the world. This usually leads to unnecessary travel which adds to extra flight costs and burden on the environment in the form of carbon emissions.

What it does

Our platform factors in organizers’ preferred locations and the participants’ locations, and recommends the optimal location to host the event, which has the lowest flight costs and emissions. The organizer can also generate and send QR codes for the participants to reduce the paper trail in these events

How does this work

● The organiser creates a conference giving the following inputs: ○ Name of the Conference. ○ Earliest Start Date of the event. ○ Latest End Date of the event. ○ Deadline of Registration. ○ Location preferences to organise the event. ● Participants explore their personalised list of events and register for the event giving their locations. ● After the deadline of registration, the organiser finalises a location and sends out emails to the participants which includes event details and uniquely generated QR-codes for participants

FEATURES

OPTIMIZATION ON FLIGHT COSTS AND EMISSIONS For every event, we fetch the real-time costs and distance of flight travel from the participant’s location to each of the organiser’s preferred locations. According to European Environmental Agency the Emission factor per passenger-kilometer in g CO2 equivalent is 285.(Source) Based on this data, we calculate the total emissions taking all the participants into account. Also factoring in the flight cost, we recommend an optimal list of locations to the organiser.

For every new participant that registers for the event: We fetch the realtime cost of flight travel from the participant location to each of the organiser’s preferred locations using the Hipmunk API. We fetch the distance of flight travel from the participant location to each of the organiser’s preferred locations.This distance of multiplied with the emission factor per passenger-kilometer in g CO2 equivalents. According to European Environmental Agency the Emission factor per passenger-kilometer in g CO2 equivalent is 285.(Source: https://ecpr.eu/Filestore/PaperProposal/daa72e5e-47be-4c2f-ae97-e8ee76cb7829.pdf) The cost and emissions are added to the values of the cost and emissions of all the other participants. The total cost and emissions for each location become the parameter to the pareto frontier module.A pareto-frontier is a set of choices that optimizes a system. Example:Assume the organiser provided 4 preferred locations. The input to the pareto-frontier module will be: graph=[ [sum of costs for location1,sum of emissions for location1], [sum of costs for location2,sum of emissions for location2], [sum of costs for location3,sum of emissions for location3], [sum of costs for location4,sum of emissions for location4], ] Since we want to minimize both the parameters,we optimize the pareto-frontier in the bottom-left direction as shown in the image below.We get an optimal list of locations in a sorted order.

QR-CODES The organizer can generate QR codes for the participants to facilitate the registration process at the event. Once the organiser finalises a location, they can send out mails.To send the mails we used python modules email and smtplib.Using the smtp python module mails can be sent to any client with a S.M.T.P listener daemon. The mails also contain unique qr-codes which are generated using the pyqrcode module.To ensure that the qr codes are unique,the qr string used to generate the qr-code was a concatenation of various parameters related to the participation registration for the event. qrstring=conference_id+participant_name+participant_email+participant_location.

PERSONALISED RECOMMENDATIONS We recommend conferences to participants taking into account their previously attended conferences.We used a dataset of 350+ conferences to train the model using the term frequency-inverse document frequency algorithm of information retrieval. According to the algorithm,a weight is assigned according to the frequency of the word in the dataset. If the word occurs more frequently it is given a higher score asserting its importance. Similar concept is applied to the location of the event. After assigning the weight to each word they are modeled in a 2D vector space and based on their closeness (angle between the words - as the angle decreases the weight should increase so the cosine_similarity function is used) the event is recommended.

What we learned

Exploring various external apis,react and firebase platforms.

What's next for Econf.ai

Recommend suitable timelines to organize the event based on the participation. ● Extend the application to accommodate local transportation and last-mile connectivity. ● Suggest complete travel itineraries for the participants attending conferences

Share this project:

Updates