Italian Dialect Survey

Come riferisci ad un gruppo di ragazzi nel tuo dialetto?

Ragazzi
Uaglioni
Tosi
Contrari
Gnari
Piccioccheddusu
Matoc
Raga'
Ragàs
Cinni
Citti
Freghi
Picciotti
Pütèi
Matèi
Boce
Bocia
Butei
Scet
Fioei
Pischelli

Come diresti la frase ‘Ho capito’?

Ho capito
Capiscivi
Go capio
So capit
Gho capìt
So capit
Ho capit'
Á laj capi
Ho inteso
Ho capi
Ho capiho
Capiscetti'
Agg' capit
A gó capì
U capì
Ó capìt
Go capit'
Lu capi
Appu cumprendiu

Come chiameresti una sedia nel tuo dialetto?

Sedia
Seggiola
Carega
Scrania
Scrana
Seggia
Hcagna
Scragna
Cadrega
Segg
Carira

Come chiameresti una sigaretta?

Sigaretta
Sigarett'
Sicaretta
Higareta
Paglia
Cicca
Ciospa
Biu
Paja
Paja
Paina
Zigarru
Zigheret
Cicchino

Come chiameresti un tiro di una sigaretta nel tuo dialetto?

Tiro
Tir'
Tirata
Na tirata
Bocada
An tir
Nu surk
Lancio
Peo

Come pronunci la parola aspetta?

Aspetta
Aspè
'Spett
Asc-pett
Aspèta
Speta
'Spetta
Peta
Hpeta
Aspetta
Aspitt

Come dici la parola ‘Gatto’ in dialetto?

Gatto
Gat
Gato
Gattu
Gatte
Gjat
Jattu
Mucill
Micio
Miscio
Gattud
Ciùcio
‘attu
Ciat

Come dici la parola ‘Cane’??

Cane
Can
Cana
Cuane
Cani
Cacciuttiell
Cà
Chène
Kene
Miscio
Gattud
Ciùcio
U Cane
‘attu

Come si chiama il salame dalle tue parti?

Salame
Salàdo
Salam
Salama
Salamu
Suppizzata
Sartizzu
Salami
Luganega
Halam
Sasezz

Come chiameresti una ‘Mucca’ nel tuo regione d’Italia?

Mucca
Vaca
Vacche
Vacca
Vaca
‘acha
Bàca
Bàga
Àca
Bacca

Come chiameresti alla spezia ‘Prezzemolo’ nel tuo dialetto?

Prezzemolo
Petrusinu
Parsemul
Petrosino
Petrusin’
Parsimolo
Prunsemu
Ptrsin
I parsû
Parsemul
Erbetta
Erbuccio
Erbuni
Pederhem
Pardesum
Pitrusinu
Petrusín'
Perdesü
Pedersèm

This interactive survey uses Machine Learning to determine an Italian speakers native region. Our motivation for building this survey was to explore the possibilities that Machine Learning could have in the study of dialects and our results prove that Machine Learning is not only accurate but also has a vast amount of applications in dialectology.

This project was developed at Fordham University under the supervision of Dr. D. Frank Hsu at the Labratory of Informatics and Data Mining

Special thanks to Dr. Erez Yoeli of Yale University for guidance throughout the development of this survey.

We developed an 11-question survey in Italian to accurately identify a respondents’ region of origin. The survey questions are mostly based on in-person surveys conducted in the summer of 2017, and supplemented with Twitter data, that was geocoded and mined for words that were most predictive of a Tweet’s origin. In an 80-person pilot survey, the survey’s out of sample AUC average was 88%. The following Jupyter Notebook created by Sasha Paulovich displays how accurate our algorithms are using cross validation along our entire data set.

Watch Alon Yoeli's YouTube documentary about Italian Dialects!

As dialects begin to fade amongst the younger generations of Italians, it is important that we learn as much as possible about regional Italian dialects for their preservation. There is no better tool for this than machine learning, especially as the volume and variety of available data grows. Now that computational processing is cheaper and more powerful than ever, it's becoming possible to quickly and automatically produce models that can analyze bigger, more complex data and deliver faster, more accurate results – on a very large scale.