We’ve been working very hard over the last few months to improve our product, make our clients happy, and get ready to launch the public beta, so we are really very happy for the prize, and thankful for the opportunity.

EverWrite Wins!

Congrats to the folks at Grupo RBS, other companies that have been part of the competition, and also the other winners, Ledface and Igluu.  Check out the coverage on TheNextWeb here. Thanks everybody!

Common Crawl is a non-profit foundation dedicated to building and maintaining an open crawl of the web, thereby enabling a new wave of innovation, education and research. Web crawl data can be used to spot trends and identify patterns in politics, economics, health, popular culture and many other aspects of life.  It provides an immensely rich corpus for scientific research, technological advancement, and innovative new businesses.  It is crucial for our information-based society that the web be openly accessible to anyone who desires to utilize it.

It’s an amazing initiative, and we’re sure that it will foster innovation in the search market. For years, Google, Microsoft and even minor search players never intended to give access to their data playground because it could allow companies to leverage this huge amount of information and become potential competitors. Now, thanks to CommonCrawl, students, researchers and startups are able to manipulate this data in order to generate value from the open web.

It’s obvious that now more than ever that you’ll need to leverage the cloud in order to benefit from this information. However, EC2 costs are always falling, and Hadoop is a continuously growing community. We wish Common Crawl a lot of success, and we are very anxious to see some innovative use cases from the Common Crawl data by the community. Have you seen something? Just let us know! Soon we’ll be writing about our experiments here, too! viagra kamagra

AOL Search Data – 20million search terms from more than 605,000 users

In 2006, the AOL search data leak was the release of detailed search logs from a large number of AOL users. The release was intentional and intended for research purposes; however, the public release meant that anyone could see the results rather than a select number of academics. AOL did not redact any information, causing privacy concerns as users could potentially be identified by their searches. This collection is awesome to use for personalization, query reformulation, or other types of research.

Yahoo! Search Query Logs for Nine Languages - 1,000 most popular searchs for 9 languages

This data set contains the 1,000 most frequent web search queries issued to Yahoo! Search for nine different languages, in 2009. The languages covered are Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, and Spanish. The data set may be useful for various information retrieval and data mining research investigations, especially those involving cross-lingual or multilingual search tasks.

Yahoo! Search Queries - 4,496 queries posted to Yahoo!’s U.S. search engine

This data set contains a random sample of 4,496 queries posted to Yahoo!’s U.S. search engine in January 2009. For privacy reasons, the query set contains only queries that have been asked by at least three different users and contain only letters of the English alphabet, sequences of numbers not longer than four numbers and punctuation characters. The query set does not contain user information and does not preserve temporal aspects of the query log.

Google NGRAMS – 1,176,470,663 five-word sequences, 13,588,391 unique words

In 2006 Google released its N-gram datasets, which it uses for developing tools such as machine translation, speech recognition, spelling correction, entity detection, information extraction, and more.

Yahoo N-GRAMS – 126 million unique sentences, 3.4 billion unique words

This dataset contains N-grams (contiguous sets of words of size n, where n = 1 to 5) extracted from a corpus of 14.6 million documents (126 million unique sentences, 3.4 billion running words) crawled from over 12,000 news-oriented sites. The data set does not contain the documents themselves but only the n-grams that occur at least twice. It provides statistics such as frequency of occurrence, number, and entropy of different left (right) single-token contexts of each n-gram. This data set may be used by researchers to build statistical language models for speech or handwriting recognition or machine translation, just like Google’s Ngrams

Yahoo! Answers Manner Questions – 142,627 questions and their answers

This data is a subset of the Yahoo! Answers corpus from a 10/25/2007 dump. It is a small subset of the questions, selected for their linguistic properties (For example, they all start with “how {to|do|did|does|can|would|could|should}”). Additionally, we removed questions and answers of obvious low quality; for example, we kept only questions and answers that have at least four words, out of which at least one is a noun and at least one is a verb. This data set may be used by researchers to learn and validate answer-extraction models for manner questions.

Yahoo! Answers Question Types – 142,627 questions and their classification

This data set contains URLs of questions posted to Yahoo! Answers, along with the question types assigned to these questions by human judges. The question types are “informational,” “advice,” “opinion,” and “polling,” which are useful for anyone interested in developing question classification systems.

Yahoo! Search queries that share clicked URLs – http://webscope.sandbox.yahoo.com/

This dataset consists of Yahoo! search queries that share clicked URLs with TREC queries. Queries that share clicked URLs are often referred to as co-clicked queries. This dataset is widely used within the information retrieval community for various Web search-related experiments.

Obscene words database – around 5,000 obscene keywords

As the guys from infochimps said, there are about 5,000 different ways to shock your grandmom in this dataset. But you can also do a lot of things. You could use this dirty-word list to screen for spammers and griefers (although there are more efficient methods publicly available), to monitor user feedback on a website, or just to improve your vocabulary. You could also use this to classify “adult search queries.” You should definitely take a look at this set.

Stop words list – 429 words that won’t help your classifier so much

This stopword list is probably the most widely used stopword list. It covers a wide number of stopwords without getting too aggressive and including too many words that a user might include in a search query. You can use this 429-word word list to remove irrelevant terms from your text analysis.

Help us!

We’d love to have your help in expanding this list, so you can contribute in the comments, and we’ll update it here. Also, we’re almost ready to launch our first open-source project, OpenKeywords (I think that, by the name, you could guess what it will do.). And BTW , did we mention that we’re hiring? Drop us a line if you’re interested.

Last week we started writing about opening our company in the U.S. from Brazil. A lot of people seemed to like the post and it went to the cover of Hacker News!

So this week we’re back our journey. On the first episode, we wrote about incorporation and filling 83(b) elections, but now we’ll be talking about obtaining an employer identification number (#EIN), a U.S. address, a U.S. phone number and a U.S. bank account.

So, let’s move on. Right after we incorporated our company, the first task was to obtain a U.S. address and phone number so we could be reached by U.S. customers and receive business-related correspondence.

Obtaining a U.S. Mailing Address:

We followed a tip from the folks at Freshdesk and opened an account at VirtualPostMail. It takes less than three minutes, and at the end of the process you have an address, probably similar to ours:

340 S LEMON AVE #3221WALNUT, CA 91789

They receive your mail, open it, scan it and make it available 24/7 online. They are an official Commercial Mail Receiving Agent (CMRA), authorized by the United States Postal Service to process and handle mail on your behalf. For us it’s working like a charm. Pricing starts at U.S. $9.99, with a three-month trial for $5.00 per month.

Obtaining a U.S. Phone Number:

Next, we started the process to get a California phone number, and we received a tip from the folks at @Deskmetrics to use a Skype online number.  We just logged into Skype and were able to choose the country and state. We paid U.S. $19.00 for three months and received our phone number: +1(415)513-0121. It was very simple and easy, and the number works 24/7.

Obtaining a #EIN (Employer Identification Number):

An Employer Identification Number (EIN) is also known as a Federal Tax Identification Number, and is used to identify a business in the U.S. The first thing you need to do is to take a look at the form SS4, and then call this number: (800) 829-4933.

The IRS Philadelphia is the only organization that processes EINs for U.S. businesses that are incorporated in foreign countries. We chatted with the attendant and filled in the form by phone, and this looks like the simplest way to do it, because you don’t have to post anything and you’ll receive your EIN code in real time, by phone. A few weeks later you’ll also receive mail confirming your EIN code and the credentials to enroll in EFTPS (the Electronic Federal Tax Payment System). After that we were able to proceed to opening our bank account.

Opening a Bank Account:

Once again we went with the recommendation from the guys at FreshDesk, and chose Silicon Valley Bank. We went in via their website and very soon we were able to begin the process.  We made some phone calls to their team, and had to make a pitch in order to start the account opening process. We found this very interesting.  In order to get the account opened, we needed to provide the following information:

• EIN;
• Scan/copy of filed/stamped articles of incorporation;
• Ownership information and founders’ passports;
• Bank Depositor Agreement (signature card) – This form identifies the contracting officers and signers on the account.
• SVBeConnect Enrollment Form – This form enrolls the account for online banking.
• Client Information Record (CIR) – Basic information about the signer(s) on the account.
• W8 Federal Form – Required for signer(s) who do not have US Social Security Number.
• U.S. Address.

In less than a week after we sent the forms, our account was open and ready for wire transfers. We also needed help from an advisor that lives in U.S. since a U.S. address of someone that is related to the company is needed.

In order to obtain credit cards, the credit limit needs to be secured with a deposit.

They have been very supportive and we’re really happy with the choice.

Next steps:

In the next post we’ll write about starting up at https://www.eftps.gov/eftps/ with EFTPS (the Electronic Federal Tax Payment System) and getting an accountant in the U.S. to help us with financial and taxes issues. If you have any tips or questions, just leave a comment. See ya’ll soon.

When we started @EverWrite, we knew we should focus on a global market and were insecure about incorporating in Brazil, due to high taxes, immature payment infrastructure, and a complicated incorporation law. We also needed to be able to establish ourselves in the U.S. in the short term.

Another goal was to open a company that would allow us not only to operate in the U.S market primarily, but also to expand internationally, so the ability to collect payments from all over the world and easily land U.S. investors was of primary importance.

When we decided to open our company in the U.S. we had already done a lot of research using resources like Quora, StartupLawyer.com, some friends, and some experienced angels and VC’s that we were lucky to meet along the journey.

This blog post from the guys at FreshDesk was also very helpful.

So we decided to write this post to share how it went for us.

When we started, we had a few premises:

• We needed to be Delaware C Corp, due to the mature Delaware corporation law.
• We wanted four-year vesting and a one-year cliff for all founders.
• We needed help and information with the 83(b) election, in order to avoid high annual taxes in case of liquidity.
• We wanted to obtain all the intellectual property and confidentiality agreements in order to be ready for fundraising.

After evaluating a few attorney candidates (most of them came through Quora), we started working with Ryan Roberts from startuplawyer.com (@startuplawyer).

As soon as we made the payment through Paypal, we received a letter of engagement and a form to http://varley.net/online/ fill out information in order to begin.

Some information that was needed:

• Name of the company
• Names of directors/founders
• Founders’ addresses
• Board president and secretary names
• Number of shares
• Shares split between founders

Once we sent that, the process started. In six days, we were able to obtain our certificate of incorporation. After 10 days, we received the full incorporation pack with:

• Bylaws
• Confidentiality and IP agreements
• Certificate of incorporation
• Stock purchase agreements
• Stock certificates
• Joint action

 The costs for incorporation were:

Lawyer and incorporation: around 2000.00 USD;

Filling Fees: around 150 USD;

On 83(b) election: This was a little tricky, but very simple to solve.

If you’re an individual who owns equity on a U.S. corporation and plan to move to the United States to become an individual taxpayer, you need to fill out an 83(b) election. However, this is not necessary if you plan to continue on in your country.

If you plan to continue on in your country and pay taxes over capital gains there, you as an individual are only subject to this kind of taxation in your country.

If you need the 83(b) election in the case you do not have an SSN or ITIN, you should fill out the form with “awaiting ITIN” in the ITIN field. Once you fill out the form, you should mail it to the IRS. Remember to register your mail. You must pay attention; you must do this no later than 30 days after the share purchase date (probably the same date as incorporation).

Right now we are still working with Ryan on the following tasks:

• Improving our Terms of Service
• Creating a stock options plan

In our next posts, we’ll discuss opening a bank account from outside the United States, obtaining a virtual address, obtaining an American phone number, and obtaining an employer identification number (#EIN). If you have any doubts, corrections or suggestions, feel free to comment below. We hope you will join us on this saga. viagra

We’re very happy to be among “Brazil’s top startups” together with @deskmetrics, a company that we also admire a lot. Right now we’re currently working on our next release (100% self serving, and accepting international credit cards) and we had to open the company and our bank accounts in the U.S. in order to sign contracts with payment gateways, banks and so on.

We’d like to thank @startuplawyer and @svb_financial for the support during this journey!

For those who asked, will be keeping our #sanpedrovalley offices in Belo Horizonte.

It started on Saturday morning, and we went through the day with pizza, beer and Faith No More. We created a (soon-to-be) open source project called OpenKeywords. The goal of the project is to be the world’s first open keyword database, allowing anyone who is interested in this kind of information to research and download the dataset for free. By now, anyone who has a Google Analytics account is able to collaborate and share their keywords with the world. On the roadmap we have other features, such as a JavaScript  tag for Web site owners and a toolbar, so expect more news soon. Right now, you can take a look at our first post on OpenKeyword’s blog.

Some cool stuff that we are using in this project:

• Ruby on Rails
• MongoDB
• Devise
• Mongoid
• Garb – Google Analytics API Wrapper
• Beer!