Juri Marcucci
@JuriMarcucci
Followers
752
Following
10K
Media
84
Statuses
559
Economist at the Bank of Italy working on Big data, Machine Learning and NLP applied to economics & finance. Views are my own. RT = I want to keep track of it.
Joined December 2012
🎓 Don't miss this discussion on how #text+#AI unlock the hidden signals in global economic strategy. 🗓️ Jun 20.🕚 11AM EDT | 5PM CEST.🎙️ Moderator: Nitish Sinha (Fed Board).🔗 Register: #EconTwitter #LLM #NLP #TradePolicy #Geoeconomics #DataScience 7/7.
lnkd.in
This link will take you to a page that’s not on LinkedIn
1
1
0
🌍 What is geoeconomic pressure?. It's when governments use #trade, #finance or #supplychains to push #geopolitical goals—with or without formal action. Think #tariffs, #sanctions export controls. or just the threat of them. #EconTwitter #AMLEDS #Webinar #LLM #AI #ML #NLP.2/7.
1
0
0
📢 Can threats alone move markets?. Join us Friday, June 20 @ 11 AM EDT / 5 PM CEST for a #AMLEDS #Webinar with @acoppola4 (@StanfordGSB) presenting:. 🧾 Geoeconomic Pressure. 🔗 #EconTwitter #LLM #NLP #ML #Text #Economics #AI.1/7
2
3
5
If you're working on:.✅ Demand estimation.✅ Consumer choice.✅ AI & economics.✅ Product differentiation.…you won’t want to miss this. #EconTwitter #AI #MachineLearning #DeepLearning #TextMining #NLP #Econometrics #DataScience #AmazonData #ConsumerBehavior #Webinar #AMLEDS.
0
0
0
🧑🏫 Moderated by @LarsNesheim (@ucl ).📄 Paper:.📍 Free & online.🗓️ Friday, May 23.⏰ 11 AM EDT | 5 PM CEST.🔗 Registration Link #econtwitter #AMLEDS.
lnkd.in
This link will take you to a page that’s not on LinkedIn
1
0
0
🔄 This is demand estimation—reimagined:.No need for hand-coded features or perfectly labeled attributes. Just smart use of **neural networks + choice data**. 📈 Scalable, low-cost, and **surprisingly accurate**. #econtwitter.
1
0
0
✨ Key Findings:.* Text & image embeddings beat standard models in predicting demand.* Embeddings from reviews give the best counterfactual performance.* Identifies close substitutes across 40 Amazon categories.* Ideal for markets with missing or hard-to-quantify product data.
1
0
0
🧠 This paper (w/ Ilya Morozov&Stephan Seiler) shows how to use pre-trained deep learning models to extract product info from images, descriptions, & reviews—then estimate demand without traditional attributes. 📦Think: Amazon-scale markets with no clean metadata. #econtwitter.
1
0
0
🚨 Can #AI estimate demand from just #text and #images?.Join us **Friday, May 23 @ 11 AM EDT / 5 PM CEST** for a cutting-edge #AMLEDS #Webinar with @GioCompiani *(@ChicagoBooth):.📊 "Demand Estimation with Text and Image Data". #econtwitter #machinelearning .🧵👇
1
1
6
📍 Online – Free & Open.🗓️ Friday, April 18.🕚 11 AM EDT | 4 PM BST | 5 PM CEST.🔗 Register here: #EconTwitter #AMLEDS.
lnkd.in
This link will take you to a page that’s not on LinkedIn
1
0
0
📄 Full paper: 🎯 If you care about #monetarypolicy, #forwardguidance, #NLP, #ML, or #AI, this webinar is for you!. #econtwitter #AMLEDS.
1
0
0
✨ Key Insights:. Text shocks explain 4x more variation in fed funds futures than rate changes. 2x stronger effects on real interest rates. Produce responses aligned with #textbook #macro models, unlike traditional surprise measures. #EconTwitter #AMLEDS.
1
0
0
🧠 Moderated by @td_econ (U. of Maryland), Amy shows that the language of #FOMC statements drives market expectations 4x more than rate changes! Using #neuralnetworks (XLNet), she builds a novel “text shock” series capturing pure forward guidance. 📉📈.#EconTwitter #AMLEDS.
1
0
1
📢 AMLEDS Webinar – Friday, April 18. 🕚 11 AM EDT / 5 PM CEST. 💬 Can the #Fed move markets with #words alone? Join us for a groundbreaking #AMLEDS webinar with @AmyHandlan (@Brown_Economics), presenting her paper: "#Text #Shocks and #Monetary #Surprises". #EconTwitter
1
1
6
🚨 Reminder: Tomorrow at 11 AM EDT / 3 PM GMT / 4 PM CET! 🚨. Join us for the @AMLEDS1 #Webinar with @SophiaKazinnik from @Stanford on "Simulating the Survey of Professional Forecasters." 📊. 🔍 #AI outperforms human predictions in forecasting! Don’t miss out!. #econtwitter #ML
0
1
2